CN112137832A - Learning system, rehabilitation support system, method, program, and learning completion model - Google Patents

Abstract

The present invention relates to a learning system, a rehabilitation auxiliary system, a method, a program and a learning completion model. The acquisition unit acquires the first rehabilitation data including at least assistant data indicating the training assistant and index data indicating the degree of recovery of the trainer regarding the rehabilitation performed by the trainer using the rehabilitation assistance system, and classifies the training by cluster analysis. The classification results obtained by the assistant. The learning unit generates a learning model inputted with second rehabilitation data including at least action data indicating an auxiliary action performed by the training assistant for the purpose of assisting the trainer, and outputs action data indicating the training assistant's next action. The learning unit generates a learning model using the second rehabilitation data preprocessed based on the classification result as teaching data.

Description

Learning system, rehabilitation assistance system, method, procedure and learning completion model

technical field

The present disclosure relates to a learning system, a rehabilitation assistance system, a method, a program, and a learning completion model.

Background technique

When a trainer such as a patient performs a rehabilitation exercise (rehabilitation), a rehabilitation assistance system such as a walking training device may be used. As an example of a walking training device, Japanese Patent No. 6052234 discloses a walking training device including a walking assistance device which is worn on the leg of the trainer to assist the trainer's walking.

When the trainer is undergoing rehabilitation, depending on the rehabilitation assistance system, training staff such as doctors and physical therapists may accompany the trainer, talk to the trainer, offer assistance, and set the rehabilitation assistance system to assist the trainer. .

However, in order to obtain good training results, it is necessary for the training staff to set and operate the rehabilitation assistance system so as to achieve proper assistance to the trainer through the rehabilitation assistance system. In addition, the timing of the setting operation, that is, the timing of addition or subtraction of assistance, and the timing of changing the degree of assistance also affects training results. Therefore, in order to perform such a setting operation, the training staff needs to determine what kind of assistance should be given to the trainer, and to determine the appropriate degree of assistance and timing. In addition, the training staff needs to judge when to talk to the trainer and when to help the trainer.

However, the current situation is that training staff make the above-mentioned judgments based on intuition and know-how, and since each training staff member has different years of experience and proficiency, training results vary greatly depending on the training staff member. Therefore, it is desired to perform appropriate assistance such that a good training result can be obtained regardless of the training staff. Therefore, in the rehabilitation support system, it is desired to suggest that the above-mentioned judgment can be made in the same way as the case where the training staff who are excellent (the evaluation of the training result is high) can assist the training staff regardless of the training staff. Technology. In addition, the assistance to the trainer is not limited to being performed by the training staff, and it is also conceivable that other types of training assistants such as human assistants are performed. Therefore, in the rehabilitation assistance system, there is a demand for a technique that can be suggested in a way that the above-mentioned judgment can be performed in the same manner as in the case of an excellent training assistant assisting, regardless of the training assistant.

SUMMARY OF THE INVENTION

The present disclosure has been made to solve such a problem, and provides a learning system or the like that generates a learning model that can suggest a preferred action to a training assistant assisting a trainer when the trainer performs rehabilitation using the rehabilitation assisting system. In addition, the present disclosure also provides a learning system or the like that generates a learning model that can suggest a preferred action to a training assistant who assists a trainer when the training is performed using the training assistance system.

The learning system according to the first aspect of the present disclosure includes an acquisition unit that acquires assistant data for at least including a training assistant who assists the trainer about the rehabilitation exercise performed by the trainer using the rehabilitation support system, indicating the above A classification result obtained by classifying the above-mentioned training assistant by cluster analysis on the action data of the auxiliary action performed by the training assistant for the purpose of assisting the above-mentioned trainer, and the first rehabilitation data representing the index data of the degree of recovery of the above-mentioned trainer; and a learning unit to generate a learning model, the learning model is input with the second rehabilitation data including at least the above-mentioned action data, to output the above-mentioned action data for instructing the next action of the above-mentioned training assistant, the above-mentioned learning section will be based on the above-mentioned classification result. The above-mentioned second rehabilitation data subjected to preprocessing is used as teaching data to generate the above-mentioned learning model. This makes it possible to generate a learning model that can suggest a preferred action to a training assistant assisting the trainer when the trainer performs rehabilitation using the rehabilitation assisting system.

It may also be characterized in that the second rehabilitation data includes at least one of the index data and the assistant data. Thereby, the index data or the assistant data can be reflected in the learned model.

It may also be characterized in that the learning unit generates the learning model by using, as teaching data, the second rehabilitation data corresponding to the training assistant included in one group in the classification result. This makes it possible to generate a learned model that takes into account the actions of the training assistants belonging to one group.

Alternatively, it may be characterized in that the learning unit uses, as teaching data, the second rehabilitation data associated with a plurality of groups tagged based on the classification result and the assistant data corresponding to each of the plurality of groups, to generate the above learning model. This makes it possible to generate a learned model in which the actions of the training assistants are considered in groups.

It may be further characterized in that the learning system includes an analysis unit that performs the cluster analysis on the first rehabilitation data to classify the training assistants, and the obtaining unit obtains a classification obtained by classifying the training assistants from the analysis unit. result. Thus, the learning system can proceed from the analysis stage.

It may be characterized in that the first rehabilitation data and the second rehabilitation data include trainer data indicating the characteristics of the trainer. Thereby, the characteristics of the trainer can be reflected in the learned model.

It may also be characterized in that the trainer data includes symptom data indicating at least one of a disease and a symptom of the trainer. Thereby, the symptom data can be reflected in the learned model.

It may also be characterized in that the action data includes at least one of data indicating an operation for changing the set value in the rehabilitation assistance system and data indicating an assisting action for the trainer. Thereby, the state of the setting value changing operation or the assisting operation can be reflected in the learned model.

It can also be characterized in that the data indicating the above-mentioned operation includes data indicating the proficiency of the above-mentioned operation. This makes it possible to reflect the proficiency of the operation in the learned model.

It may also be characterized in that the learning unit generates the learning model by using the second rehabilitation data corresponding to the training assistant included in the group as teaching data for each of the plurality of groups in the classification result. Thereby, a plurality of learned models can be generated.

The learning system may be further characterized in that the learning system includes a group specifying unit for specifying a group in the classification result, and the learning unit specifies the second complex corresponding to the training assistant included in the group specified by the group specifying unit. The health data is used as teaching data to generate the above-mentioned learning model. Thereby, the learned model of only the designated group can be generated.

A learning system according to a second aspect of the present disclosure includes an acquisition unit that acquires assistant data indicating a training assistant who assists the trainer, including at least the training performed by the trainer using the training assistance system, and indicating that the training assistant A classification result obtained by classifying the above-mentioned training assistants by cluster analysis on the action data of the auxiliary action performed for the purpose of assisting the above-mentioned trainer, and the first data of the index data indicating the degree of improvement of the physical function of the above-mentioned trainer; and a learning unit , generating a learning model that is input with second data including at least the above-mentioned action data, to output the above-mentioned action data for instructing the next action of the above-mentioned training assistant, and the above-mentioned learning unit will be based on the above-mentioned classification result. The above-mentioned second data is used as teaching data to generate the above-mentioned learning model. This makes it possible to generate a learning model that can suggest a preferred action to a training assistant assisting the trainer when the trainer executes the training using the training assistance system.

A rehabilitation assistance system according to a third aspect of the present disclosure is a rehabilitation assistance system capable of accessing a learned model, which is a learned model learned by the learning system according to the first aspect, and includes an output unit that communicates with a trainer The second rehabilitation data related to the rehabilitation exercise performed using the rehabilitation assistance system is output as an input to the learned model; and a notification unit notifies the action data output from the learned model to the above-mentioned model. The above-mentioned training assistant assisting the above-mentioned trainer in the rehabilitation exercise. Thereby, when the trainer performs rehabilitation using the rehabilitation assistance system, it is possible to suggest a preferred action to the training assistant who assists the exercise.

It may also be characterized in that the rehabilitation assistance system includes a designation unit for specifying the training assistant who assists the trainer in the rehabilitation exercise, the rehabilitation assistance system can access a classification result storage unit that stores the classification result, and when When the training assistant designated by the designation unit is a training assistant that did not use the teaching data when generating the learned model, the output unit outputs the second rehabilitation data, and the notification unit notifies the training assistant. As a result, unnecessary notifications are not performed for training assistants that are supposed to not require notification.

A learning method according to a fourth aspect of the present disclosure includes an acquisition step of acquiring assistant data indicating a training assistant who assists the trainer, including at least the rehabilitation exercise performed by the trainer using the rehabilitation support system, indicating the above A classification result obtained by classifying the above-mentioned training assistant by cluster analysis on the action data of the auxiliary action performed by the training assistant for the purpose of assisting the above-mentioned trainer, and the first rehabilitation data indicating the index data of the above-mentioned degree of recovery of the above-mentioned trainer; and The learning step generates a learning model, the learning model is inputted with the second rehabilitation data including at least the above-mentioned action data, to output the above-mentioned action data for instructing the next action of the above-mentioned training assistant, and the above-mentioned learning step will be based on the above-mentioned classification results. The above-mentioned second rehabilitation data subjected to preprocessing is used as teaching data to generate the above-mentioned learning model. This makes it possible to generate a learning model that can suggest a preferred action to a training assistant assisting the trainer when the trainer performs rehabilitation using the rehabilitation assisting system.

A rehabilitation assistance method (operation method of a rehabilitation assistance system) according to a fifth aspect of the present disclosure is a rehabilitation assistance system capable of accessing a learned model that is a learned model learned by the learning method according to the fourth aspect The rehabilitation assistance method, comprising: an output step, wherein the rehabilitation assistance system outputs the second rehabilitation data related to the rehabilitation exercise performed by the trainer using the rehabilitation assistance system as an input to the learned model. and a notification step, wherein the above-mentioned rehabilitation assistant system notifies the above-mentioned training assistant who assists the above-mentioned trainer in the above-mentioned rehabilitation exercise of the above-mentioned action data output from the above-mentioned learned model. Thereby, when the trainer performs rehabilitation using the rehabilitation assistance system, it is possible to suggest a preferred action to the training assistant assisting the exercise.

A program according to a sixth aspect of the present disclosure is a program for causing a computer to execute a step of acquiring a data indicating that the exerciser is supported, including at least the rehabilitation exercise performed by the exerciser using the rehabilitation assistance system. The assistant data of the training assistant, the action data indicating the auxiliary action performed by the training assistant for the purpose of assisting the trainer, and the first rehabilitation data indicating the index data of the degree of recovery of the trainer are classified by cluster analysis. A classification result obtained by training an assistant; and a learning step of generating a learning model that is input into the second rehabilitation data including at least the above-mentioned action data, to output the above-mentioned action data for enlightening the next action of the above-mentioned training assistant, The said learning process generates the said learning model by using the said 2nd rehabilitation data preprocessed based on the said classification result as teaching data. This makes it possible to generate a learning model that can suggest a preferred action to a training assistant assisting the trainer when the trainer performs rehabilitation using the rehabilitation assisting system.

A rehabilitation assistance program according to a seventh aspect of the present disclosure is a rehabilitation program for causing a computer that can access a rehabilitation assistance system of a learned model, which is a learned model obtained by using the program of the sixth aspect, to execute the following steps Auxiliary program: an output step of outputting the above-mentioned second rehabilitation data related to the rehabilitation exercise performed by the trainer using the above-mentioned rehabilitation auxiliary system as an input to the above-mentioned learning completed model; and a notification step, from the above-mentioned learning completed model. The above-mentioned action data output by the model is notified to the above-mentioned training assistant who assists the above-mentioned trainer in the above-mentioned rehabilitation exercise. Thereby, when the trainer performs rehabilitation using the rehabilitation assistance system, it is possible to suggest a preferred action to the training assistant who assists the exercise.

The learned model according to the eighth aspect of the present disclosure is a learning model obtained by learning using the learning system according to the first (or second) aspect, a learning model obtained by learning with the learning method according to the fourth aspect, and any one of the learning models obtained by using the program learning according to the sixth aspect. Thereby, when a trainer performs rehabilitation (or training) with the rehabilitation assistance system (or training assistance system), it is possible to provide a learned model that can suggest a preferred action to a training assistant assisting the exercise.

According to the present disclosure, it is possible to provide a learning system that generates a learning model that can suggest a preferred action to a training assistant assisting a trainer when performing rehabilitation using the rehabilitation assisting system. In addition, according to the present disclosure, a rehabilitation assistance system using the generated learned model, a method and program for learning the learned model, a learned model, and a rehabilitation assistance method and program using the learned model can be provided. In addition, the present disclosure can also be applied to training other than rehabilitation, whereby the same effect can be obtained for training other than rehabilitation.

The above and other objects, features and advantages of the present disclosure will be more fully understood from the following detailed description and the accompanying drawings, which are given by way of illustration only and should not be considered limiting of the present disclosure.

Description of drawings

FIG. 1 is an overall schematic diagram showing a configuration example of a rehabilitation assistance system according to Embodiment 1. FIG.

FIG. 2 is a schematic perspective view showing a configuration example of a walking assistance device in the rehabilitation assistance system of FIG. 1 .

FIG. 3 is a block diagram showing an example of a system configuration of a walking training device in the rehabilitation assistance system of FIG. 1 .

FIG. 4 is a block diagram showing a configuration example of a server in the rehabilitation support system of FIG. 1 .

FIG. 5 is a flowchart for explaining an example of learning processing in the server of FIG. 4 .

FIG. 6 is a flowchart for explaining an example of rehabilitation assistance processing in the server of FIG. 4 .

FIG. 7 is a diagram showing an example of an image presented to the training staff in the rehabilitation support process of FIG. 6 .

FIG. 8 is a diagram showing an example of an image presented to the training staff in the rehabilitation support process of FIG. 6 .

9 is a block diagram showing a configuration example of a server in the rehabilitation support system according to Embodiment 2. FIG.

FIG. 10 is a schematic diagram showing an example of the result of the cluster analysis performed on the server of FIG. 9 .

FIG. 11 is a flowchart for explaining an example of learning processing in the server of FIG. 9 .

Detailed ways

Hereinafter, the present disclosure will be described based on the embodiments of the invention, but the inventions according to the claims are not limited to the following embodiments. In addition, it is not limited that all the structures demonstrated in the embodiment are necessary as means for solving the problem.

<Embodiment 1>

Hereinafter, Embodiment 1 will be described with reference to the drawings.

(System Components)

FIG. 1 is an overall schematic diagram showing a configuration example of a rehabilitation assistance system according to Embodiment 1. FIG. The rehabilitation assistance system (rehabilitation system) according to the present embodiment mainly includes a walking training device 100 , an external communication device 300 , and a server (server device) 500 .

The walking training device 100 is a specific example of a rehabilitation assistance device that assists the exercise (rehabilitation exercise) of the trainer (user) 900 . The walking training device 100 is a device for a hemiplegic patient who is paralyzed in one leg, that is, a trainer 900 to perform walking training according to the instruction of the training staff 901 . Here, the training staff 901 can be a therapist (physical therapist) or a physician, and can also be referred to as a training instructor, a training assistant, a training assistant, or the like because he assists the training of the trainer by instructing, assisting, or the like. As exemplified here, training staff 901 is a human.

The walking training device 100 mainly includes: a control panel 133 attached to the frame 130 constituting the overall skeleton; a treadmill 131 for the trainee 900 to walk; The sick leg.

The frame 130 is erected on the treadmill 131 installed on the floor. The treadmill 131 rotates the endless belt 132 by a motor (not shown). The treadmill 131 is a device for promoting the walking of the trainer 900 , and the trainer 900 who performs the walking training climbs on the belt 132 and attempts a walking action in accordance with the movement of the belt 132 . In addition, for example, as shown in FIG. 1 , the trainer 901 can also stand on the belt 132 behind the trainer 900 and perform the walking action together, but it is generally preferable that the trainer 900 be easily performed by standing astride the belt 132 or the like. state of help.

The frame 130 supports the control panel 133 housing the overall control unit 210 for controlling motors and sensors, and a training monitor 138 serving as a liquid crystal panel for notifying the trainer 900 of the progress of the training. In addition, the frame 130 supports the front stretcher 135 near the upper part of the head of the trainer 900, the protective belt stretcher 112 near the top of the head, and the rear stretcher 137 near the back of the upper head. Additionally, the frame 130 includes a handrail 130a for the trainer 900 to grasp.

The armrests 130a are arranged on the left and right sides of the trainer 900 . Each handrail 130a is arranged along the direction parallel to the walking direction of the trainee 900 . The armrest 130a can adjust the up-down position and the left-right position. That is, the armrest 130a can include a mechanism for changing its height and width. Moreover, the handrail 130a can also be comprised so that the inclination angle may be changed, for example by adjusting so that the front side and the rear side of a walking direction may differ. For example, the handrail 130a can have an inclination angle that gradually increases in the walking direction.

Moreover, the armrest sensor 218 which detects the load received from the trainee 900 is provided in the armrest 130a. For example, the armrest sensor 218 can be a load detection sheet of a resistance change detection type in which electrodes are arranged in a matrix. In addition, the armrest sensor 218 may be a 6-axis sensor in which a 3-axis acceleration sensor (x, y, z) and a 3-axis gyro sensor (roll, pitch, yaw) are combined. However, the type and installation position of the armrest sensor 218 are arbitrary.

The camera 140 functions as an imaging unit for observing the whole body of the trainee 900 . The camera 140 is installed in the vicinity of the training monitor 138 so as to face the trainer. The camera 140 captures a still image and a moving image of the trainee 900 under training. The camera 140 includes a set of a lens and an imaging element such that the angle of view of the whole body of the trainee 900 can be captured. The imaging element is, for example, a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor, and converts the optical image formed on the imaging surface into an image signal.

Through the cooperative operation of the front side stretching portion 135 and the rear side stretching portion 137, the load of the walking assistance device 120 is offset so that the load of the walking assistance device 120 does not become a burden on the sick leg, and the sick leg is assisted according to the set degree. Swing action.

One end of the front side wire 134 is connected to the winding mechanism of the front side stretching portion 135 , and the other end is connected to the walking assistance device 120 . The take-up mechanism of the front-side stretching portion 135 winds or draws out the front-side wire 134 according to the movement of the diseased leg by turning on/off a motor (not shown). Similarly, one end of the rear wire 136 is connected to the winding mechanism of the rear pulling part 137 , and the other end is connected to the walking assistance device 120 . The winding mechanism of the rear side pulling part 137 winds up or draws out the rear side wire 136 according to the movement of the diseased leg by turning a motor (not shown) on/off. Through the cooperative operation of the front stretching portion 135 and the rear stretching portion 137, the load of the walking assistance device 120 is offset so that the load does not become a burden on the sick leg, and the patient is assisted according to the set degree. The swinging action of the legs.

For example, the training staff 901 as an operator sets a high level of assistance to a severely paralyzed trainer. If the level at which the assistance is performed is set to be large, the front side stretching portion 135 winds up the front side wire 134 with a relatively large force in accordance with the swing timing of the sick leg. If the training progresses without assistance, the training staff 901 sets the level of assistance to be minimal. When the level at which the assistance is performed is set to the minimum, the front side pulling part 135 winds the front side wire 134 with the force of only canceling the self-weight of the walking assistance device 120 in accordance with the swing timing of the sick leg.

The walking training device 100 includes a fall prevention belt device as a safety device including a harness 110 , a protective belt wire 111 , and a protective belt stretching portion 112 as main components. The harness 110 is a belt wound around the abdomen of the trainer 900, and is fixed to the waist by, for example, a surface fastener. The harness 110 is provided with the connecting hook 110a which connects one end of the protective tape wire 111 which is a sling, and can also be called a suspender. The trainer 900 wears the harness 110 so that the connection hook 110a is located on the back.

One end of the protective tape wire 111 is connected to the connecting hook 110 a of the back strap 110 , and the other end is connected to the winding mechanism of the protective tape stretching portion 112 . The winding mechanism of the protective tape stretching portion 112 winds up or draws out the protective tape wire 111 by turning on/off a motor not shown. With this structure, when the trainer 900 is about to fall, the fall-prevention protective belt device winds up the protective belt wire 111 according to the instruction of the overall control unit 210 that has detected the movement, and supports the upper body of the trainer 900 by the harness 110 . Prevents trainer 900 from falling.

The harness 110 includes a posture sensor 217 for detecting the posture of the trainee 900 . The posture sensor 217 is, for example, a combination of a gyro sensor and an acceleration sensor, and outputs the inclination angle of the abdomen on which the harness 110 is worn with respect to the direction of gravity.

The management monitor 139 is attached to the frame 130 and is a display input device mainly used for monitoring and operation by the training staff 901 . The management monitor 139 is, for example, a liquid crystal panel, and a touch panel is provided on the surface thereof. The management monitor 139 displays various menu items related to training settings, various parameter values during training, training results, and the like. Moreover, the emergency stop button 232 is provided in the vicinity of the monitor 139 for management. When the training worker 901 presses the emergency stop button 232, the walking training apparatus 100 is stopped urgently.

The walking assistance device 120 is worn on the sick leg of the trainer 900 , and assists the trainer 900 in walking by reducing the load of extension and flexion at the knee joint of the sick leg. The walking assistance device 120 includes a sensor or the like that measures the load on the sole of the foot, and outputs various data related to the moving leg to the overall control unit 210 . In addition, the harness 110 can also be connected to the walking assist device 120 using a connection member (hereinafter, referred to as a hip joint) having a rotating portion. Details of the walking assistance device 120 will be described later.

The overall control unit 210 generates rehabilitation data that may include setting parameters related to training settings, various data related to moving legs output from the walking assistance device 120 as a training result, and the like. The rehabilitation data can include data indicating the training worker 901 or the number of years of experience, proficiency, etc., data indicating the symptoms, walking ability, recovery, etc. of the trainer 900 , sensors provided outside the walking assistance device 120 , and the like. Various output data, etc. However, the details of the rehabilitation data will be described later.

The external communication device 300 is a specific example of transmission means that transmits rehabilitation data to the outside. The external communication device 300 can have a function of temporarily storing the rehabilitation data output by the walking training device 100 and a function of transmitting the stored rehabilitation data to the server 500 .

The external communication device 300 is connected to the control panel 133 of the walking training device 100 via, for example, a USB (Universal Serial Bus) cable. In addition, the external communication device 300 is connected to a network 400 such as the Internet or a local area network via a wireless communication device 410 , for example, through a wireless LAN (Local Area Network). In addition, the walking training device 100 may further include a communication device instead of the external communication device 300 .

The server 500 is a specific example of a storage means for storing rehabilitation data. The server 500 is connected to the network 400 and has a function of accumulating the rehabilitation data received from the external communication device 300 . The function of the server 500 will be described later.

In the first embodiment, the walking training device 100 is described as an example of the rehabilitation assisting device, but the present invention is not limited to this, and may be a walking training device of another configuration, or may be a rehabilitation assisting exerciser. of any rehabilitation aids. For example, the rehabilitation assisting device may be an upper limb rehabilitation assisting device assisting the rehabilitation of shoulders and arms. Alternatively, the rehabilitation assisting device may be a rehabilitation assisting device for assisting the exerciser's balance ability.

Next, the walking assistance device 120 will be described using FIG. 2 . FIG. 2 is a schematic perspective view showing a configuration example of the walking assistance device 120 . The walking assistance device 120 mainly includes a control unit 121 , a plurality of frames supporting each part of the sick leg, and a load sensor 222 for detecting a load applied to the sole of the foot.

The control unit 121 includes an assist control unit 220 that controls the walking assist device 120 and a motor (not shown) that generates a driving force for assisting the extension and flexion motion of the knee joint. The frame supporting each part of the sick leg includes a thigh frame 122 and a lower leg frame 123 that is rotatably connected to the thigh frame 122 . The frame further includes a sole frame 124 rotatably connected to the calf frame 123 , a front connecting frame 127 for connecting the front wire 134 , and a rear connecting frame 128 for connecting the rear wire 136 .

The thigh frame 122 and the lower leg frame 123 rotate relative to each other about the hinge axis _Ha shown in the figure. The motor of the control unit 121 rotates according to the instruction of the auxiliary control unit 220, and applies force so that the upper leg frame 122 and the lower leg frame 123 are relatively opened or closed around the hinge axis _Ha . The angle sensor 223 housed in the control unit 121 is, for example, _a rotary encoder, and detects the angle formed by the thigh frame 122 and the lower leg frame 123 around the hinge axis Ha. The calf frame 123 and the sole frame 124 rotate relative to each other around the illustrated hinge axis _Hb . The angular range of relative rotation is adjusted in advance by the adjustment mechanism 126 .

The front side connecting frame 127 is provided to extend in the left-right direction on the front side of the thigh, and is connected to the thigh frame 122 at both ends. Moreover, in the front side connection frame 127, the connection hook 127a for connecting the front side wire 134 is provided in the center vicinity of the left-right direction. The rear side connecting frame 128 is provided to extend in the left-right direction on the rear side of the lower leg, and is connected to the lower leg frame 123 extending vertically at both ends, respectively. Moreover, in the rear side connection frame 128, the connection hook 128a for connecting the rear side wire 136 is provided in the center vicinity of the left-right direction.

The thigh frame 122 includes a thigh strap 129 . The thigh belt 129 is a belt provided integrally with the thigh frame, and is wound around the thigh of the diseased leg to fix the thigh frame 122 to the thigh. Thereby, the whole of the walking assistance device 120 is prevented from being displaced with respect to the leg of the trainer 900 .

The load sensor 222 is a load sensor embedded in the sole frame 124 . The load sensor 222 detects the magnitude and distribution of the vertical load on the soles of the feet of the trainer 900 , and can also be configured to detect COP (Center Of Pressure: center of load), for example. The load sensor 222 is, for example, a load detection sheet of a resistance change detection type in which electrodes are arranged in a matrix.

Next, an example of a system configuration of the walking training device 100 will be described with reference to FIG. 3 . FIG. 3 is a block diagram showing an example of a system configuration of the walking training apparatus 100 . As shown in FIG. 3 , the walking training device 100 can include an overall control unit 210 , a treadmill drive unit 211 , an operation reception unit 212 , a display control unit 213 , and a stretch drive unit 214 . In addition, the walking training device 100 can include a belt drive unit 215 , an image processing unit 216 , a posture sensor 217 , an armrest sensor 218 , a communication connection IF (interface) 219 , an input/output unit 231 , and the walking assistance device 120 .

The overall control unit 210 is, for example, an MPU (Micro Processing Unit), and executes control of the entire apparatus by executing a control program read from a system memory. The overall control unit 210 can include a walking evaluation unit 210a, a training determination unit 210b, an input/output control unit 210c, and a notification control unit 210d, which will be described later.

The treadmill drive unit 211 includes a motor that rotates the belt 132 and a drive circuit thereof. The overall control unit 210 executes the rotation control of the belt 132 by sending a drive signal to the treadmill drive unit 211 . The overall control unit 210 adjusts the rotational speed of the belt 132 in accordance with, for example, the walking speed set by the training worker 901 .

The operation accepting unit 212 accepts an input operation from the training staff 901 and transmits an operation signal to the overall control unit 210 . The training staff 901 operates the operation buttons provided in the device, the touch panel superimposed on the management monitor 139 , the attached remote control, and the like, which constitute the operation accepting unit 212 . This operation enables on/off of the power supply, an instruction to start training, input of numerical values related to settings, and selection of menu items. In addition, the operation accepting unit 212 can also accept an input operation from the trainer 900 .

The display control unit 213 receives the display signal from the overall control unit 210 , generates a display image, and displays it on the training monitor 138 or the management monitor 139 . The display control unit 213 generates an image showing the progress of the training and a live image captured by the camera 140 based on the display signal.

The stretching driving part 214 includes a motor for stretching the front wire 134 and a driving circuit thereof, which constitute the front stretching part 135, and a motor for stretching the rear wire 136, which constitutes the rear stretching part 137, and its driving circuit. Drive circuit. The overall control unit 210 controls the winding of the front wire 134 and the winding of the rear wire 136 by sending a driving signal to the stretching driving unit 214, respectively. In addition, not only the coiling operation, but also the stretching force of each wire is controlled by controlling the driving torque of the motor. The overall control unit 210, for example, determines the timing at which the diseased leg is switched from the standing leg state to the swinging leg state based on the detection result of the load sensor 222, and assists the swinging action of the diseased leg by increasing or decreasing the stretching force of each wire in synchronization with the timing. .

The protective tape driving portion 215 includes a motor for stretching the protective tape wire 111 and a driving circuit thereof, which constitute the protective tape stretching portion 112 . The overall control unit 210 controls the winding of the protective tape wire 111 and the stretching force of the protective tape wire 111 by sending a driving signal to the protective tape driving unit 215 . For example, when the trainer 900 is predicted to fall, the overall control unit 210 winds up a certain amount of the protective band wire 111 to prevent the trainer from falling.

The image processing unit 216 is connected to the camera 140 and can receive image signals from the camera 140 . The image processing unit 216 receives an image signal from the camera 140 in accordance with an instruction from the overall control unit 210 , and performs image processing on the received image signal to generate image data. In addition, the image processing unit 216 can perform image processing on the image signal received from the camera 140 in accordance with an instruction from the overall control unit 210 to perform specific image analysis. For example, the image processing unit 216 detects the position (leg standing position) of the foot of the diseased leg in contact with the treadmill 131 by image analysis. Specifically, the leg standing position is calculated by, for example, extracting an image area near the distal end of the sole frame 124 and analyzing the identification mark drawn on the belt 132 overlapping the distal end.

The posture sensor 217 detects the inclination angle of the abdomen of the exerciser 900 with respect to the gravitational direction as described above, and transmits a detection signal to the overall control unit 210 . The overall control unit 210 uses the detection signal from the posture sensor 217 to calculate the posture of the trainer 900 , specifically, the inclination angle of the trunk. Among them, the overall control unit 210 and the posture sensor 217 may be connected by wire or by short-range wireless communication.

The armrest sensor 218 detects the load applied to the armrest 130a. That is, the load of the amount which the trainer 900 cannot fully support his own body weight with both legs is applied to the armrest 130a. The armrest sensor 218 detects this load and transmits a detection signal to the overall control unit 210 .

The overall control unit 210 also functions as a function execution unit that executes various calculations and controls related to control. The walking evaluation unit 210a evaluates whether or not the walking motion of the trainer 900 is abnormal walking using data acquired from various sensors. The training determination unit 210b determines a training result for a series of walking training based on, for example, the cumulative number of abnormal walking evaluated by the walking evaluation unit 210a. The overall control unit 210 can generate the determination result, the cumulative number of abnormal walking that is the root of the determination result, and the like as a part of the rehabilitation data.

However, the method of determination is arbitrary including the criterion of this determination. For example, the motion amount of the paralyzed body can be compared with a reference for each walking phase and can be determined. Among them, the walking phase classifies one walking cycle (one walking cycle) about the sick leg (or healthy leg) into a leg standing period in the standing leg state, and a transition period from the leg standing period to the leg swing period in the leg swing state. , swing leg period, transition period from swing leg period to leg standing period, etc. For example, as described above, it is possible to classify (determine) which walking phase is in based on the detection result of the load sensor 222 . In addition, the walking cycle can be one cycle of the leg standing period, the transition period, the leg swing period, and the transition period as described above, but which period is defined as the start period is arbitrary. In addition to this, the walking cycle can be, for example, a two-leg support state, a single-leg (sick leg) support state, a two-leg support state, and a single-leg (healthy leg) support state as one cycle. In this case, the Which state is defined as the start state is also arbitrary.

In addition, the walking cycle focusing on the right leg or the left leg (healthy leg or diseased leg) can be further subdivided. For example, the leg standing period can be divided into initial grounding and 4 periods, and the leg swing period can be divided into 3 periods. to express. The initial grounding refers to the moment when the foot is grounded on the floor, and the four stages of the leg-stance period are the load response period, the mid-leg-stance phase, the end-leg-stance phase, and the forward-swing leg phase. The load response period is the period from the initial grounding to the moment when the foot on the opposite side leaves the floor (the opposite side leaves the ground). The mid-leg standing period is the period from the time when the opposite side leaves the ground to the moment when the heel of the foot is observed to leave the ground (the heel leaves the ground). The end of leg stance is the period from the heel off the ground to the initial grounding on the opposite side. The forward leg swing period is the period from the initial ground contact on the opposite side to the time when the foot is observed to leave the floor (lift off the ground). Stage 3 of the leg swing period refers to the early stage of the leg swing, the middle stage of the leg swing, and the late stage of the leg swing. The initial stage of the leg swing is the period from the end of the forward leg swing period (the above-mentioned lifting off the ground) to the time when the feet are crossed (the feet are crossed). The mid-leg swing is the period from when the feet cross until the tibia becomes vertical (tibia is vertical). The end of the swing is the period from the vertical of the tibia to the next initial touchdown.

The communication link IF 219 is an interface connected to the overall control unit 210 , and is an interface for giving commands to the walking assistance device 120 worn on the sick leg of the trainee 900 and receiving sensor information.

The walking assistance device 120 can include a communication link IF229 that is wired or wirelessly connected to the communication link IF219. The communication connection IF 229 is connected to the assist control unit 220 of the walking assist device 120 . The communication connection IFs 219 and 229 are communication interfaces such as wired LAN or wireless LAN conforming to a communication standard.

In addition, the walking assist device 120 can include an assist control unit 220 , a joint drive unit 221 , a load sensor 222 , and an angle sensor 223 . The assist control unit 220 is, for example, an MPU, and executes control of the walking assist device 120 by executing a control program in accordance with an instruction from the overall control unit 210 . In addition, the assistance control unit 220 notifies the overall control unit 210 of the state of the walking assistance device 120 via the communication links IFs 219 and 229 . In addition, the assist control unit 220 receives an instruction from the overall control unit 210 and executes control such as start/stop of the walking assist device 120 .

The joint drive part 221 includes the motor of the control unit 121 and its drive circuit. The auxiliary control unit 220 sends _a drive signal to the joint drive unit 221 to apply force so that the thigh frame 122 and the lower leg frame 123 are relatively opened or closed about the hinge axis Ha. Through such an action, the extension action and flexion action of the knee are assisted, and the knee is prevented from being bent.

The load sensor 222 detects the magnitude and distribution of the vertical load on the soles of the feet of the trainer 900 as described above, and transmits a detection signal to the assist control unit 220 . The assist control unit 220 receives and analyzes the detection signal to perform state determination, switching estimation, and the like of swinging/standing legs.

The angle sensor 223 detects the angle formed by the thigh frame 122 and the lower leg frame 123 around the hinge axis Ha as described above, and transmits _a detection signal to the assist control unit 220 . The assist control unit 220 receives the detection signal and calculates the opening angle of the knee joint.

The input/output unit 231 includes, for example, a USB (Universal Serial Bus) interface, and is a communication interface for connecting to external devices (the external communication device 300 and other external devices). The input/output control unit 210c of the overall control unit 210 communicates with an external device via the input/output unit 231, and performs the above-mentioned rewriting of the control program in the overall control unit 210 and the control program in the auxiliary control unit 220, acceptance of commands, and creation of The output of rehabilitation data, etc. The walking training device 100 communicates with the server 500 via the input/output unit 231 and the external communication device 300 under the control of the input/output control unit 210c. For example, the input/output control unit 210 c can perform control of transmitting rehabilitation data to the server 500 via the input/output unit 231 and the external communication device 300 and control of receiving commands from the server 500 .

When notification to the training worker 901 is required, the notification control unit 210d controls the display control unit 213 or a separately provided sound control unit or the like to notify from the management monitor 139 or a separately provided speaker. The details of this notification will be described later, but the case where notification is required to the training staff 901 can be a case where an instruction for notification is received from the server 500 .

Next, the details of the server 500 will be described.

As described above, the walking training device 100 transmits various types of rehabilitation data to the server 500 via the external communication device 300 . The server 500 can be configured to receive rehabilitation data from a plurality of walking training devices 100, and thus can collect a large amount of rehabilitation data. Also, the server 500 is a processing device that processes various data. For example, the server 500 can function as a learning device (learning system) that performs machine learning using the collected rehabilitation data and constructs a learned model. The learning device can also be a learner. In addition, the learning apparatus can also be referred to as a learning model generating apparatus.

FIG. 4 is a block diagram showing a configuration example of the server 500 . As shown in FIG. 4 , the server 500 can include a control unit 510 , a communication IF 514 , a data storage unit 520 , and a model storage unit 521 . The control unit 510 is, for example, an MPU, and controls the server 500 by executing a control program read from the system memory. The control unit 510 can include a level determination unit 510a, a learning unit 510b, and a response processing unit 510c, which will be described later. In this case, the control program described above includes a program for realizing the functions of these parts 510a to 510c.

The communication IF 514 includes, for example, a wired LAN interface, and is a communication interface for connecting to the network 400 . The control unit 510 can receive rehabilitation data from the walking training device 100 via the communication IF 514 , and can transmit a command to the walking training device 100 .

The data storage unit 520 includes, for example, a storage device such as an HDD (hard disk drive) and an SSD (solid state drive), and stores rehabilitation data. The control unit 510 writes the rehabilitation data received from the external communication device 300 via the communication IF 514 into the data storage unit 520 .

The model storage unit 521 also includes storage devices such as HDD and SSD. In addition, the data storage unit 520 and the model storage unit 521 may have a common storage device. The model storage unit 521 stores at least one of a learning model that has not been learned (including the case of learning) (hereinafter, referred to as an unlearned model) and a learned learning model (hereinafter, referred to as a learned model). When the server 500 functions as a learning device, at least an unlearned model is stored in the model storage unit 521 . When the server 500 cooperates with the walking training device 100 to execute the rehabilitation assistance process, the model storage unit 521 stores at least an applicable learned model.

In addition, the control unit 510 can be configured to perform control to switch between the function as a learning device and the function of performing rehabilitation assistance processing using the learned model. However, the server 500 can also be distributed according to the apparatuses used in the learning phase and the apparatuses used in the operation phase accompanied by the learned model. The level determination unit 510a and the learning unit 510b are provided to make the server 500 function as a learning device, and the response processing unit 510c is provided to cause the server 500 to execute a part of the rehabilitation support process.

(rehabilitation data)

Here, before describing the level determination unit 510a, the learning unit 510b, and the response processing unit 510c, rehabilitation data that can be collected by the server 500 for learning or rehabilitation assistance processing will be described. The rehabilitation data that can be collected by the server 500 mainly includes: (1) the setting parameters of the walking training device 100 , (2) detection data detected by sensors installed in the walking training device 100 , and (3) related to the trainer 900 data, (4) data related to the training staff 901. The rehabilitation data of the above (1) to (4) can be collected in association with the acquisition time and date. In addition, detection data or setting parameters may be collected as log data in time series, or may be feature quantities extracted from data at regular intervals.

The rehabilitation data is mainly data obtained by operation input, automatic input, measurement of sensors, and the like in the walking training device 100 . In addition, the rehabilitation data can also include video data recorded by the camera 140 . In addition, the rehabilitation data may be data for each implementation day of rehabilitation, and in this case, may also be referred to as daily data. Hereinafter, the collection of the rehabilitation data generated by the walking training device 100 will be described by the server 500 , but the server 500 may be configured to acquire a part of the rehabilitation data from, for example, another server other than the walking training device 100 . Part of the rehabilitation data referred to here can be, for example, the detailed data of the above (3) such as the symptoms of the trainer 900, the detailed data of the above (4) such as the number of years of PT experience, and the like. The former can be stored in another server as the medical record information of the trainer 900, and the latter can be stored and stored in another server as the PT's resume or the like.

In the learning phase, the server 500 only needs to receive the rehabilitation data from the walking training device 100 periodically at the time of generation of the rehabilitation data, every day, every week, or the like. In the learning phase and the operation phase, the type of rehabilitation data to be used (content included in the rehabilitation data) can be made different. For example, in the operation phase, the server 500 only needs to receive the rehabilitation data from the walking training device 100 at the start of training, and only need to receive the data with changes in the above (1) and (2) during the training. In addition, any one of the walking training device 100 and the server 500 may serve as the main body to perform transmission and reception of rehabilitation data.

The above (1) will be described.

The data of the above (1) can be defined together with the detection data of the above (2) as the training data of the trainer 900 acquired by the walking training apparatus 100 during the rehabilitation implementation.

The setting parameters of the walking training device 100 are, for example, data input by an operator or automatically set data in order to set the operation of the walking training device 100 . However, as described above, the operator is usually the training staff 901 who actually accompanies the training of the trainer 900 , and the following description assumes that the operator is the training staff 901 . In addition, since the training staff 901 is often a physical therapist (PT: Physical Therapist), the training staff 901 may be abbreviated as "PT" hereinafter.

In the walking training device 100, the difficulty of walking training can be adjusted by setting parameters. However, the setting parameter may further include a parameter indicating the level of difficulty, and in this case, a part or all of the other setting parameters may be changed along with the change of the level. As the recovery of the trainer 900 progresses, the training staff 901 gradually increases the difficulty of the walking training. That is, as the walking ability of the trainer 900 becomes higher, the training staff 901 reduces the assistance of the walking training device 100 . In addition, when an abnormality is seen in the walking training, the training staff 901 adds assistance. By appropriately adjusting the setting parameters by the training staff 901 , the trainer 900 can perform appropriate walking training and can perform rehabilitation more efficiently.

Specific examples of the setting parameters are shown below.

The setting parameters include, for example, partial body weight free weight [%], vertical position of armrest 130a [cm], horizontal position of armrest 130a [cm], presence or absence of hip joint, ankle plantar flexion restriction [deg] , Ankle dorsiflexion limit [deg] and so on. In addition, as setting parameters, for example, treadmill speed [km/h], swing assist [horizontal], and swing front-to-back ratio [front/rear] can also be mentioned. In addition, as setting parameters, for example, knee extension assistance [horizontal], knee flexion angle [deg], knee extension time [sec], auxiliary height [mm], weight loss threshold [%], load can also be mentioned. Threshold [%]. In addition, the setting parameters include, for example, the inclination [degree] of the belt of the treadmill, the assistance [horizontal] of the movement of the joint by the walking assist device, the assist of the movement of the joint by the walking assist device or the swing assist. Frequency, abnormal or normal walking judgment conditions (for example, judgment thresholds), fall or fall-like judgment conditions (for example, judgment thresholds), and when reports are made in accordance with abnormal or normal walking, its occurrence conditions (occurrence frequency, generation threshold, etc.). Here, the report may be based on any one of sound, vibration, display, etc., and may include a part or all of it. In addition, the unit of the data included in the rehabilitation data is arbitrary, including the setting parameters exemplified here.

The partial body weight free weight is the ratio of the weight of the trainer 900 to be freed by the protective belt stretching portion 112 pulling the protective belt wire 111 . The higher the difficulty level of the desired walking training, the lower the value of the partial body weight free load is set by the training staff 901 . The up-down position and the left-right position of the armrest 130a are adjustment amounts from the reference position of the armrest 130a. The presence or absence of the hip joint refers to whether or not the hip joint is attached. The ankle joint plantar flexion limit and the ankle joint dorsiflexion limit define the angular range within which the calf frame 123 and the sole frame 124 can rotate about the hinge axis _Hb . The ankle plantarflexion limit corresponds to the upper anterior angle, and the ankle dorsiflexion limit corresponds to the posterior maximum angle. That is, the ankle joint plantar flexion restriction and the ankle joint dorsiflexion restriction are respectively restriction values for the angle of bending the ankle joint to the side where the toe is lowered and the side where the toe is raised. The training staff 901 sets the values of the ankle joint plantar flexion limit and the ankle joint dorsiflexion limit so that the higher the difficulty of the desired walking training, the larger the angle range.

The treadmill speed is based on the walking speed of the treadmill 131 . The higher the difficulty level of the desired walking training, the higher the treadmill speed is set by the training staff 901 . The swing assist is a degree corresponding to the stretching force given by the front wire 134 when the legs are swung, and the higher the degree, the greater the maximum stretching force. The higher the difficulty level of the desired walking training, the lower the level of the swing assist is set by the training staff 901 . The swing front-to-back ratio is the ratio of the stretching force of the front wire 134 to the stretching force of the rear wire 136 at the time of the swing of the leg.

The knee extension assist is applied to the extent corresponding to the drive torque of the joint drive unit 221 to prevent the knee from being bent when standing up, and the higher the extent, the greater the drive torque. The higher the difficulty level of the desired walking training, the lower the degree of knee extension assistance is set by the training staff 901 . The knee flexion angle is the angle at which knee extension assistance is performed. The knee flexion and extension time is a period during which knee extension assistance is performed. If the value is large, the knee is flexed and extended slowly, and if the value is small, the knee is assisted in rapid flexion and extension.

The auxiliary height is the height of a component such as a cushion provided on the sole of the leg opposite to the paralyzed leg of the trainer 900 (the leg on the side not wearing the assist device, that is, the leg on the side of the walking assist device 120 ). The weight loss threshold is one of the thresholds of the load applied to the sole of the foot, and when the threshold is lower than the threshold, the swing assist is released. The load threshold is one of the thresholds of the load applied to the sole of the foot, and when the threshold is exceeded, the swing assist is performed. In this way, the walking assistance device 120 can be configured to adjust the bending and extending motion of the knee according to the four setting parameters of the knee flexion angle, the knee flexion and extension time, the weight loss threshold, and the load threshold.

In addition, the walking training device 100 may be configured to feed back, for example, the set value, target value, target achievement rate, target value of various parameters such as load and angle to the trainer and/or training staff through a speaker (not shown), for example. realization time, etc. The above-mentioned setting parameters can also include parameters related to the presence or absence of such feedback sound, the volume, and the like.

Besides, the above-mentioned setting parameters may not be directly related to training. For example, the above-mentioned setting parameters may be set for the training monitor 138 or the not-shown speaker to provide images, music, the type of game, the difficulty of the game, and the like in order to increase the motivation of the trainer 900 . value etc.

In addition, the above-mentioned setting parameters are an example, and setting parameters other than these may exist. Alternatively, some of the above-mentioned setting parameters may not exist. In addition, as described above, the above-mentioned setting parameters are many parameters for adjusting the difficulty level of training, but may also include parameters irrelevant to the level of difficulty. For example, the walking training device 100 can be configured to display an icon image for attention-calling that is displayed on the training monitor 138 . Further, as setting parameters irrelevant to the difficulty level, for example, parameters for improving the concentration of the trainer 900 on training, such as the size of the icon image for attracting attention, the display interval, and the like can be exemplified. In addition, time information such as the date and time when the setting operation was completed, or timing information other than time (for example, information indicating the difference between the leg standing period and the leg swing period in one walking cycle) can be added to the above-mentioned setting parameters in advance. .

The above (2) will be described.

The detection data of the above (2) can be defined together with the data of the above (1) as the training data of the trainer 900 acquired by the walking training apparatus 100 during the rehabilitation implementation.

The detection data mainly includes sensor data. The sensor data are sensor values detected by various sensors of the walking training apparatus 100 . For example, the sensor data are the inclination angle of the trunk detected by the posture sensor 217 , the load detected by the armrest sensor 218 , the inclination angle, the angle detected by the angle sensor 223 , and the like. Sensors that output sensor data are an acceleration sensor, an angular velocity sensor, a position sensor, a light sensor, a torque sensor, a weight sensor, and the like. In addition, an encoder provided in a motor of the front wire 134, the rear wire 136, the winding mechanism of the protective tape wire 111, or the like can be used as a sensor. In addition, a torque sensor (a load cell) of the motor may be used as a sensor, and a current detection unit that detects a drive current value of the drive motor may be used as a sensor.

In addition, the sensor data can include, for example, line-of-sight data acquired by a line-of-sight detection sensor that detects line-of-sight. The same line of sight data can also be obtained based on taking an image of at least the eyes of the trainer 900 and detecting the line of sight through image processing, or it is also possible to determine the orientation of the face (upward/facing direction based on the image of at least the face of the trainer 900 being taken) inferior) to obtain. Such data can also be included in the above-mentioned detection data. In addition, the detection data may be audio data acquired by a voice acquisition unit such as a microphone that acquires the voice of the trainer 900 or the training staff 901 , text data obtained by analyzing the voice data, or data obtained by analyzing the text data. The voice of the training staff 901 can include a shout to the trainer 900 related to the correction of the walking style or the like. In addition, the sensor data may be data in which the brain waves of the trainer 900 are detected by an electroencephalogram, or data in which the brain waves of the training worker 901 are detected by an electroencephalogram.

In addition, the sight line detection sensor, the imaging unit that captures the above-mentioned image, the microphone, etc. can be installed on the main body side of the walking training device 100 , but can also be installed in, for example, a glasses-type wearable terminal for the trainer 900 to wear. The terminal only needs to include a wireless communication unit that wirelessly communicates data by a wireless communication method such as Bluetooth (registered trademark), and also includes a wireless communication unit on the side of the walking training apparatus 100 . Thereby, the walking training apparatus 100 can acquire the data acquired by the wearable terminal through wireless communication. The electroencephalograph is limited to an electroencephalograph with high detection accuracy, and can be configured to be installed on the main body side of the walking training apparatus 100 to detect the brainwaves of the trainer 900 and the training staff 901 by distinguishing them. Among them, it is preferable to install the electroencephalograph at a position close to the detection target person, such as the above-mentioned glasses-type wearable terminal (eg, the temple part of glasses).

In addition, the detection part which acquires detection data, such as a sensor, is not limited to the structure demonstrated with reference to FIGS. 1-3, or the structure exemplified as the glasses-type wearable terminal or the like. For example, the trainer 900 can be made to wear clothes equipped with a wearable biometric sensor and/or a wearable touch sensor. The clothing mentioned here is not limited to the clothing worn on the upper body, but may also be the clothing worn on the lower body, and may also be a set of upper and lower clothing, for example, a part of the harness 110 and other components worn. In addition, the clothes and the walking training apparatus 100 are provided with the above-mentioned wireless communication unit. Thereby, the walking training apparatus 100 can acquire data acquired by the wearable biometric sensor and the wearable touch sensor through wireless communication. Wearable biosensors can acquire vital data such as the wearer's heart rate. The wearable touch sensor can acquire data indicating that the wearer, that is, the trainer 900 is touched from the outside, that is, the information indicating the position where the trainer 901 touched the trainer 900 .

In addition, the detection data is not limited to values represented by detection signals detected by various sensors, etc., and may include values calculated based on detection signals from a plurality of sensors, statistical processing of data from one or a plurality of sensors, and the like. Statistical value of the detected signal. As the statistical value, various statistical values such as an average value, a maximum value, a minimum value, and a standard deviation value can be used, and a static statistical value may be used, for example, one day, one training session, and one walking session. Statistical value of dynamic statistics within a certain period such as cycle.

For example, the sensor data can include the opening angle of the knee joint calculated from the angle between the thigh frame 122 and the lower leg frame 123 detected by the angle sensor 223 . Also, the sensor data on the angle sensor can include an angular velocity obtained by differentiating the angle. The sensor data about the acceleration sensor may be a velocity obtained by integrating the acceleration and a position obtained by integrating the acceleration twice.

For example, the detection data can include an average value, a total value, a maximum value, a minimum value, and a representative value as follows for each exercise of rehabilitation on a daily or daily basis. Examples of the average value here include an average speed (total walking distance/total walking time) [km/h], an average value of step distances [cm], and a walking rate [steps] indicating the number of steps per minute (steps). /min], walking PCI [beats/m], fall avoidance assistance [%], etc. The average speed can be, for example, a value calculated from a speed setting value of the treadmill 131 or a value calculated from a drive signal in the treadmill drive unit 211 . The stride is the distance between the heel of one side touching the ground and the heel of the same side touching the ground again next time. PCI refers to Physiological CostIndex (a clinical index of Physiological Cost Index), and walking PCI indicates energy efficiency while walking. Fall avoidance assistance [%] refers to the number of times that the training staff 901 provided fall avoidance assistance to the trainer 900 per one step, that is, the ratio of fall avoidance assistance [times], that is, per one step. Percentage of fall avoidance assistance.

In addition, as the total value here, walking time [seconds], walking distance [m], number of steps [steps], fall avoidance assistance [times], fall avoidance assistance parts, and the number of times for each part [times], etc. .

In addition, as the maximum value or the minimum value here, the maximum value and the minimum value of the continuous walking time [second], the continuous walking distance [m], the number of continuous steps [steps], the maximum value and the minimum value of the walking PCI [beat/m] can be mentioned. The minimum value (in other words, the longest value of the walkable distance per beat), etc. As a representative value, the value most used as the speed of the treadmill 131 (representative speed [km/h]), etc. can be mentioned.

In this way, the detection data can include data directly or indirectly supplied from detection units such as various sensors. In addition, time information such as the date and time when the detection was completed, or timing information other than time can be added to the above-mentioned detection data in advance.

In addition, the above-mentioned detection data is an example, and other detection data may exist. Alternatively, some of the above-mentioned detection data may not exist. That is, when the detection data is used as the rehabilitation data, the server 500 only needs to collect one or more detection data.

The above (3) will be described.

Data related to the trainer 900 (hereinafter, referred to as trainer data) indicates, for example, attributes of the trainer 900 and the like. The trainer data can include symptom information, Br. Stage, SIAS, initial walking FIM, latest walking FIM, and the like, represented by the age, sex, and physique (height, weight, etc.) of the trainer 900 . In addition, the trainer data may include the name or ID of the trainer 900 , and may also include preference information indicating the preferences of the trainer 900 , character information indicating the character, and the like. In addition, the trainer data can include sports items other than those related to walking ability as FIM, and can also include cognitive items. That is, the trainer data can include various data representing the physical ability of the trainer 900 . Among them, a part or all of the trainer data can also be referred to as physical information, basic information, trainer characteristic information, or the like.

Here, the symptom information can include information indicating the initial symptom, its onset time, and the current symptom, and it can be understood that the trainer 900 needs rehabilitation mainly because of the symptoms included here. However, symptom information can also include symptoms not directly related to rehabilitation. In addition, the symptom information can include the part (injury part) together with the type (disease name or disease name) of the disease suffered from stroke (cerebrovascular disease) and spinal cord injury, and can include the classification according to the type. For example, stroke can be classified into cerebral infarction, intracranial hemorrhage (cerebral hemorrhage/subarachnoid hemorrhage), and the like.

Br.Stage refers to Brunnstrom Recovery Stage, aiming at the recovery process of hemiplegia. According to the observation, the recovery stage is divided into 6 stages. The trainer data can include the main item related to the walking training device 100 in the Br.stage, ie, the lower body item. SIAS stands for Stroke Impairment Assessment Set, which is an index that comprehensively evaluates stroke dysfunction. SIAS can include hip flexion test (Hip-Flex), knee extension test (Knee-Ext), foot floor test (Foot-Pat). In addition, SIAS can include lower extremity touch (TouchL/E), lower extremity position sense (PositionL/E), abdominal muscle strength (Abdominal), and verticality test (Verticality).

FIM (Functional Independence Measure: Functional Independence Scale) determines one of the evaluation methods for evaluating ADL (Activities of Daily Life). In FIM, evaluation is performed on seven stages of 1 to 7 points according to the amount of assistance.

For example, the walking FIM is a general indicator showing the degree of recovery. If a person can walk more than 50m without a helper and without a harness (assistant), it becomes the highest score of 7 points, and if a helper can only walk less than 15m, it becomes the lowest score of 1 point. In addition, when the movement of 50 m was possible with minimal assistance (the amount of assistance was 25% or less), it was awarded 4 points, and when the movement of 50 m was possible with the medium level of assistance (the amount of assistance was 25% or more), it was awarded 3 points. Thus, the walking FIM of the trainer 900 gradually increases as recovery progresses. In addition, the walking distance in the case of performing the evaluation of the walking FIM is not limited to 50 m, and, for example, may be 15 m.

From this, it can be seen that the latest walking FIM managed by the walking training device 100 is not only an index indicating the physical ability of the trainer 900 but also an index indicating the degree of recovery of the trainer 900 from the time when the rehabilitation starts. The walking FIM is an index indicating the exercise ability of the trainer 900 without using the actuator, that is, the walking ability. In other words, the walking FIM is an important indicator for knowing the progress of rehabilitation of the trainer 900 . In addition, the amount of change or the speed of change from the initial walking FIM to the latest walking FIM also serves as an index indicating the degree of recovery. The rate of change can also be referred to as FIM efficiency, and can be, for example, dividing the gain (change amount) of FIM up to now by the number of days of rehabilitation, the number of days elapsed indicating the period of rehabilitation, or when the trainer 900 is an admitted patient Values obtained during periods such as the number of days of hospitalization below.

In addition, the walking FIM can be understood as a score under evaluation conditions such as wearing an assist device, and in this case, information indicating the conditions applied during the evaluation can be added to the information indicating the walking FIM. The conditions can include settings such as the auxiliary height when the information is acquired, the harness used (for example, the walking assistance device 120, other walking assistance devices, no harness, etc.), the angle setting of the knees and ankles in the harness, etc., Walking on a flat ground or walking on an incline. In addition, the normal walking FIM is a walking FIM under level walking, and the level walking information indicating the level walking information can also include information such as the furthest walking distance (maximum continuous walking distance [m]) during the level walking evaluation.

In this way, the trainer data of (3) above can include index data including at least one of symptoms, physical ability, and degree of recovery of the trainer 900 regarding the rehabilitation performed by the trainer 900 using the walking training apparatus 100 . In addition, the data that can be included in the concepts of both the physical ability such as the latest walking FIM and the degree of recovery generally only needs to be included in one, but may be included in both. In addition, in the same case, for all the items of the rehabilitation data, the data of a certain item can be regarded as any one or a plurality of data in the above-mentioned (1) to (4). In addition, the above-mentioned trainer data can be preliminarily added with time information such as the measurement time and date of the walking FIM, such as the time and date when it was acquired.

The above (4) will be described.

Data related to the training staff 901 (hereinafter, referred to as staff data) indicates, for example, attributes of the training staff 901 and the like. The staff data includes the name, ID, age, sex, physique (height, weight, etc.) of the training staff member 901 , the name of the hospital to which they belong, the number of years of experience as a PT or a doctor, and the like. The staff data can include, as data related to the helper, a digitized value of the time spent helping the trainer 900 .

In addition, in the case of multiple training workers assisting in rehabilitation at the same time, the rehabilitation data can include data of multiple workers. In addition, each worker data may further include information indicating whether it is a main training worker or an auxiliary training worker. In addition to or instead of such information, each worker data may include a training worker indicating whether the setting operation on the management monitor 139 or the confirmation of the image is performed, or whether it is only a starter. Information and the like to the training staff who support the role of the trainer 900 by hand.

Moreover, it is preferable that the walking training apparatus 100 is comprised so that the rehabilitation plan for the trainee 900 can be input. Furthermore, the data of the rehabilitation plan input in this way can also be included as worker data related to the training worker 901 as the input person or as rehabilitation data belonging to other categories. In addition, in order to be able to cope with the change of the training staff 901 , the walking training device 100 is preferably configured so as to be capable of inputting precautions and notices when assisting the training of the trainer 900 in the future. Furthermore, the data input in this way can also be included as worker data related to the training worker 901 as the input person or as rehabilitation data belonging to other categories.

The reason why these data are included in the rehabilitation data is that there may be a case where a certain training worker can successfully carry out the training of the trainer 900 simply because there are precautions and information from other skilled training workers. In addition, time information, such as the input time and date of the rehabilitation plan, etc., can be added in advance to the above-mentioned staff data, such as the time and date of completion of the input.

(Learning phase)

Next, processing in the learning phase (learning period) of the control unit 510 of the server 500 will be described with reference to FIG. 5 . FIG. 5 is a flowchart for explaining an example of the learning process in the server 500 .

The control unit 510 performs preprocessing on a part or all of the information included in the rehabilitation data as described above, performs machine learning using the processed data, and constructs a learned model from an unlearned model. The level determination unit 510a executes preprocessing (preparatory processing), and the learning unit 510b executes machine learning. However, the control unit 510 can also be configured to execute preprocessing other than the processing in the level determination unit 510a together.

First, the control unit 510 of the server 500 prepares a plurality of sets of data for learning (actually preprocessing). Therefore, the control unit 510 prepares, for example, the first rehabilitation data collected within a predetermined period as one set of learning data. For example, the first rehabilitation data collected in one walk training or one execution of the walking training may be prepared as one set of learning data. In addition, in the following description, one set of learning data is also referred to as a data set. The first rehabilitation data is data related to rehabilitation performed while the trainer 900 uses the walking training apparatus 100 and is assisted by the training staff 901 as needed.

Among them, the one-time walking training is a series of training performed by one trainer 900 , and when the one-time walking training ends, the next trainer 900 performs training in the walking training device 100 . 1 walk training is usually about 20 to 60 minutes. One execution of the walking training is one unit of continuous walking for 900 trainers in one walking training. 1 walk training consists of multiple implementations. For example, one implementation takes about 5 minutes. Specifically, in one walking training session, the trainer 900 rests for 5 minutes after performing the walking training for 5 minutes. That is, in one walking training, the execution of the walking training and the rest are alternately repeated. The 5 minutes between the rest and the rest is the time for one implementation. Of course, the time for one training session and one execution time is not particularly limited, and can be appropriately set for each trainee 900 .

In addition, the control unit 510 may prepare first rehabilitation data collected in a period shorter than one implementation as learning data, or may prepare rehabilitation data collected in a longer period than one implementation as one set Learning data.

And the level determination part 510a inputs the 1st rehabilitation data prepared in this way (step S1). Next, the level determination part 510a determines the level which shows the evaluation (for example, the degree of excellence) of the training staff based on the inputted first rehabilitation data (step S2). The level determination unit 510a can be said to be a screening unit for screening training staff (eg, excellent training staff).

The level determination unit 510a is an example of an output unit (level output unit) that outputs the degree indicating the evaluation of the training staff, and the determination result of the level determination unit 510a is an example of an output result from the degree output unit. That is, the level can be an example of the degree, and, although not specifically described, levels related to other values can also be an example of the degree. Hereinafter, the level output unit will be described by taking the level determination unit 510a as an example. However, the degree output unit may be, for example, a part that calculates and outputs the index value based on the evaluation of the training staff as an example of the degree. For example, the level determination unit 510a can determine and output the level indicating the evaluation of the training staff based on such an index value.

The above-mentioned first rehabilitation data may be part or all of the above-mentioned rehabilitation data, and at least include a part of the staff data and a part of the index data. In other words, the first rehabilitation data corresponds to the rehabilitation data that is used in the preprocessing stage (level determination stage) of learning and includes at least staff data and index data.

As described above, the staff data is data indicating the training staff 901 assisting the trainer 900, and can include, for example, the name or ID of the training staff 901, and information indicating the hospital to which they belong. It is particularly preferred that the staff data used here contain the names or IDs used to identify the training staff 901 . As described above, the index data is data indicating the degree of recovery of the trainer 900 , and can include, for example, the FIM efficiency of the walking FIM.

The level determination unit 510a can perform determination based on a predetermined determination criterion. As the predetermined criterion, for example, from the viewpoints of FIM efficiency, walking speed, and walking stability, one or more of the following conditions (a) to (d) may be satisfied. However, the criterion for determination is not limited to this, and the number of years of experience can be cited as the simplest example. Among them, the FIM efficiency is an example of a value representing the recovery speed of the trainer.

(a) The average or maximum value of the FIM efficiency (for example, the length of the period until the FIM becomes 6 minutes or more until it becomes possible to walk without assistance) of all the trainers assisted by the target training staff is the threshold value the following.

(b) The average or minimum value of the walking speed of all the trainers assisted by the target training staff is equal to or greater than the threshold value. Alternatively, the rate of increase of the walking speed is equal to or greater than the threshold value.

(c) The average or maximum value of the frequency of abnormal walking in the flat ground walking (walking on the treadmill 131 ) of all the trainers assisted by the target training staff is equal to or less than the threshold value. Alternatively, the rate of decrease in the frequency is equal to or greater than the threshold value.

(d) The index of the graceful walking of all the trainers assisted by the target training staff is equal to or greater than the threshold value. Here, the first rehabilitation data includes an index indicating the gracefulness of walking. Alternatively, the rate of increase for this metric is above a threshold.

In each of the above (a) to (d), a threshold value group consisting of m−1 threshold values is prepared for the number of levels m. In addition, the respective threshold value groups of (a) to (d) described above are mutually different threshold value groups. In addition, in the above (a) to (d), the threshold processing is performed on the data of all the trainers assisted by the target training staff, but it is also possible to perform the threshold processing on all the rehabilitation data assisted by the target training staff. Thresholding. Accordingly, it is also conceivable that two or more training staff assist each trainer at the same time or during different periods.

In addition, threshold processing can also be performed for the rehabilitation data for distinguishing whether the training staff participates in the rehabilitation as a main staff member or as an auxiliary staff member in the rehabilitation. Similarly, it is also possible to perform threshold value processing on the rehabilitation data for distinguishing whether the training staff participates in rehabilitation as a staff member of the operation management monitor 139 or as a staff member providing assistance (hand support).

Taking a simple example, the level determination unit 510a considers that the above (a) to (d) are all a level number of 2, and obtains through threshold processing whether or not he is an excellent training worker. The training staff who are judged to be excellent are judged to be excellent (above the prescribed level). In addition, in a simpler example, the level determination unit 510a can determine an excellent job by performing threshold processing for determining whether or not an excellent training worker is an excellent training worker using only the above-mentioned (a) as the condition and 2 as the level number. personnel.

In order to make such a determination, it is basically necessary to discriminate and train workers in advance. Therefore, in order to distinguish training workers, it can be said that it is preferable that the worker data include names or IDs as described above. In addition, when the staff data does not include such information, the training staff can also be briefly distinguished by other information such as years of experience and age, for example.

It is particularly preferable that the level determination unit 510a determines the level for each characteristic of the trainer 900 . However, in this case, it is assumed that the first rehabilitation data and the second rehabilitation data to be described later include trainer data representing the characteristics of the trainer 900 . The characteristics of the trainer 900 include height, weight, sex, disease, symptoms, and the like. As a result, the level determination unit 510a can classify, for example, for each gender of the trainer 900, a trainer of the gender can be called an excellent training worker.

It is particularly preferable that the trainer data include symptom data indicating at least one of the disease (disease name or disease name) and the symptom of the trainer 900 . This is for predicting that the training staff is good or bad based on the disease or symptoms of the trainer 900 . The symptom data is data in which the above-mentioned symptom information is described. In particular, in the case of walking training, symptoms included in the symptom data include, for example, backward movement of the trunk, forward leaning of the trunk, movement of the trunk on the affected side, flexion of the knee joint, difficulty in lifting the toes, difficulty in holding the leg swing, and difficulty in keeping the trunk. Posterior tilt, pelvic retraction, lower extremity forward tilt, knee extension, knee flexion, swing. In addition, as symptoms included in the symptom data, for example, movement on the healthy side of the trunk, tiptoeing, pelvic elevation, external rotation of the hip joint, circumduction, and medial whip can also be mentioned. As a result, the level determination unit 510a can classify the trainer of the disease and symptom for each disease and symptom of the trainer 900 and can be called an excellent trainer.

In addition, the level determination unit 510a may be configured to determine the level for each value indicated by the index data such as the initial FIM of the trainer 900 . Thereby, the level determination part 510a can classify|categorize the trainer which has each value for each value shown by the index data, and can be called an excellent training worker.

The learning unit 510b generates (constructs) the second rehabilitation data corresponding to the training staff (that is, the training staff who are excellent at a certain level) as a result of the judgment in the level judgment section 510a is judged to be at a predetermined level or higher, as teaching data. Finished model. The second rehabilitation data includes at least action data indicating an auxiliary action performed by the training staff for the purpose of assisting the trainer. The learned model generated by the learning unit 510b is a model to which such second rehabilitation data is input, and action data for indicating the next action (next auxiliary action) of the training worker is output. Generation of such a learned model will be described.

Here, the type of the unlearned model to be learned by the learning unit 510b and the algorithm thereof are arbitrary, and as the algorithm, a neural network can be used, and a deep neural network (DNN) in which a hidden layer is multi-layered is particularly preferably used. As the DNN, for example, a feedforward (forward propagation type) neural network such as a multilayer perceptron (MLP) using an error back propagation method can be used. In addition, as the learning method used by the learning unit 510b in this way (the same is true for the learning method used by the learning unit described in the third embodiment), a well-known algorithm can be used, and a detailed description thereof will be omitted and briefly described.

Here, an example in which the learning unit 510b generates a learned model using MLP will be described, and examples of input parameters input to the unlearned model and output parameters output from the unlearned model in the learning unit 510b will be described. The input parameters correspond to the nodes of the input layer, respectively, and the output parameters correspond to the nodes (ie, dependent variables) of the output layer, respectively. In addition, as described above, the unlearned model is not limited to the complete unlearned case, but also includes the case of the model under learning, and the learned model refers to a model in a stage where it can be used.

As described above, the second rehabilitation data includes at least action data. That is, the input parameters input to the unlearned model include a part or all of the above-mentioned action data. Here, the item of the action data refers to an item indicating an auxiliary action. Items of the action data can be, for example, an operation indicating that a certain setting parameter is set to a certain value, an operation indicating that the setting parameter is set to another value, an action of supporting the trainer's waist with a hand, a hand Information on any of various auxiliary actions such as the action of supporting the trainer's shoulders.

Since the unlearned model and the learned model are models that output action data, the output parameters also include some or all of the action data items. In addition, since there are two or more input parameters to the unlearned model, the second rehabilitation data includes data of two or more items, and the same is true for the learned model. Of course, both the action data in the second rehabilitation data and the action data as output parameters can include items representing each of a plurality of auxiliary actions.

The action data will be described from the viewpoint of the acquisition route. The action data is a part of the detection data of the above-mentioned (2) in the above-mentioned rehabilitation data, and can include, for example, data indicating that the trainer is touched by the training staff from the outside. In addition, the action data may include the setting parameters of the above (1) set in the walking training apparatus 100 by the training worker, and data obtained by extracting the actions of the training worker from the video recording data. In addition, the setting parameters included in the action data may also include setting parameters automatically set based on default values or the like, and it is particularly preferable to include setting parameters that are automatically set in succession to the setting content at the time of the previous implementation.

As described above, the learning unit 510b generates a learned model using the second rehabilitation data corresponding to the training staff determined to be at or above the predetermined level as teaching data. Therefore, immediately after step S2, the learning part 510b selects the 2nd rehabilitation data concerning the training worker of the predetermined level or more as teaching data (step S3).

Therefore, the level determination unit 510a or the learning unit 510b can be configured to automatically assign the same correct answer label to the second rehabilitation data of the training staff of a predetermined level or higher. Alternatively, the level determination unit 510a or the learning unit 510b may be configured to automatically assign a correct answer label corresponding to the level to the second rehabilitation data of the training worker whose level is equal to or higher than a predetermined level. As an example, the second rehabilitation data of the training staff whose level is 7 or more in all 10 levels is used as the teaching data. In this case, for example, the correct answer label (correct answer variable) can be assigned to "1.0" for the second rehabilitation data related to the training worker whose best level is 10. Furthermore, for example, the correct answer variable can be assigned to "0.9", "0.8", and "0.7" for the second rehabilitation data related to each training worker at levels 9, 8, and 7, respectively. In this way, as the determined level is higher, it is possible to assign a correct answer variable with a value that contributes more to the construction of the learning model (change of weight coefficient and threshold).

In addition, the explanation has been made on the premise that the second rehabilitation data of the training staff below the predetermined level is not used for learning, but it can also be given by setting the correct answer variable of the output parameter to be the correct answer to "0" or the like. Labels that indicate incorrect answers are used. It can be said that the use of the second rehabilitation data of the training staff below the predetermined level corresponds to the use of the negative teaching data. In addition, it is also possible to assign incorrect answer labels corresponding to levels in the same way as assigning correct answer labels corresponding to levels. In the case of the above example, for example, for the second rehabilitation data related to each training worker whose level is 4 and 1, the correct answer variable that becomes the output parameter of the correct answer can be assigned to "0.4" and "0.1", respectively. . In addition, correct answer labels and the like can also be assigned manually.

Then, the learning unit 510b inputs the selected teaching data to the unlearned model, and generates a learned model (step S4). When a forward propagation type neural network such as an MLP is used, the learning unit 510b can input data sets at each time when rehabilitation starts and during rehabilitation as one data set. Among them, the learning unit 510b can input the data group counted for a predetermined period as one data group for each predetermined period. Alternatively, the learning unit 510b may input a data group counted for a predetermined period (a period longer than a unit time) from each time point as one data group for each time point. In addition, in any case, one data set may be a data set for which statistics are performed for a certain period of time such as one step and one walking cycle, and in this case, it may be a data set that starts to be input for each of the above-mentioned predetermined periods. .

When generating the learned model, the learning unit 510b inputs an appropriate number of times to the unlearned model for each of a plurality of sets of teaching data. For example, a learned model is generated using a part of the teaching data (learned training data), and the accuracy of the learned model is checked using the remaining groups as test data. As for the inspection results, if the accuracy is good, it is installed as it is, and if the accuracy is poor, the pre-processing is changed, or the learned model is generated and evaluated again after performing processing such as adjustment. In addition, both evaluation data for checking the accuracy and test data for testing the final accuracy can be prepared. In addition, it is possible to generate a learned model reflecting the item of the data set input when generating the learned model.

In addition, the hyperparameter to be adjusted is arbitrary. The above-mentioned objects include, for example, the number of layers of the neural network, the number of units (nodes) in each layer, the number of repetitions of learning using the same data set (number of epochs), and the input to the model at a time. The number of data (batch size). In addition, as the above-mentioned objects, for example, learning coefficients, types of activation functions, and the like can also be mentioned. In addition, the learning coefficient is also called a learning rate, and can be a value that determines how much to change the weight of each layer at a time.

Through the above processing, it is possible to construct a learned model that outputs action data representing auxiliary actions to be revealed based on the current state. Furthermore, each output parameter only needs to be associated with an item to be revealed in the action data, respectively. As a result, as will be described later, in the walking training device 100 using the learned model, the acquired data can be used as input parameters, and the action data indicating the auxiliary action to be revealed can be output to inform the training staff of the assistance. action.

In addition, it is preferable that the second rehabilitation data include at least one of index data and staff data. Thereby, it is possible to vary the content of the enlightenment according to the level of the training staff or according to the value of the index data of the trainer (for example, FIM efficiency, etc.).

In addition, it is preferable that the action data includes at least one of assist execution data and setting operation data. The assisting execution data is data representing assisting actions for the trainer, and data in which the trainer assists the trainer by assisting with bare hands or the like can be used as data detected from sensors, image processing, and the like.

In addition, the setting operation data is data showing the operation of changing the setting value in the walking training apparatus 100 , in other words, data showing the usage of the setting value. The setting operation data can include, for example, the time required from when the setting screen is opened on the management monitor 139 until the setting operation or all setting operations are completed. )The data. This is because it can be guessed to some extent whether the training staff is experienced or not based on the proficiency of the operation. In addition, the operation accepting unit 212 cannot determine whether the operation was performed by the training staff 901 or the trainer 900. However, if the rehabilitation is designated by the training staff 901, it is only considered that the operation is performed by the training staff 901. The operations performed by the training staff 901 may be processed. Of course, it is also possible to determine whether the operator is the training worker 901 or the trainer 900 based on the imaging data captured by the camera 140 .

As can be seen from these examples, the items included in the second rehabilitation data can be the same as the items included in the first rehabilitation data. However, the second rehabilitation data can also exclude some items such as the name and ID of the training staff from the first rehabilitation data.

Next, other kinds of learning models are exemplified. A part of the second rehabilitation data can also be input as image data to a feature extraction unit including a convolutional layer and a pooling layer in a CNN (Convolutional Neural Network). As the image data, for example, image data showing the trajectory of the COP for 10 steps, etc. can be mentioned. When such a feature extraction unit is provided, it is also possible to input the result of the feature extraction here and other input parameters to all the connection layers in parallel.

In addition, as the neural network, for example, a neural network having a recursive structure such as RNN (Recurrent Neural Network) can also be used. In addition, the RNN can also be extended into a neural network with LSTM (Long short-term memory) blocks (there is also a case where it is referred to as LSTM for short). When using a recursive model such as an RNN, one data group may include time-series data such as detection data, for example, in order for the learning unit 510b to sequentially input the second rehabilitation data at each time in one implementation. That is, one data group (learning data group) may include time-series log data. In addition, one data group may include the feature amount extracted from the log data as described above, or may include image data obtained by performing data processing on time-series detection data.

In addition, when using a recursive model such as an RNN, the learning unit 510b may input, for example, a data group counted for a predetermined period as one data group for each predetermined period. Alternatively, when a recursive model is used, the learning unit 510b may input a data group counted for a predetermined period (a period longer than a unit time) from each time as one data group for each time. In addition, one data set may be a data set for which statistics are performed for a certain period of time such as one step and one walking cycle, and in this case, it may be a data set to be inputted for each of the above-mentioned predetermined periods. In addition, the category of such statistical processing can further include the above-described processing of performing data processing on time-series detection data to obtain image data.

As a result, it is possible to construct timely output based on the current and previous past states, and to timely output only the period obtained from the period of one data group such as the above-mentioned predetermined time and the period obtained from the number of storage steps, indicating that the auxiliary action to be revealed at the present, which is predicted from the past, can be constructed. The learned model of the action data. Furthermore, as will be described later, in the walking training device 100 using the learned model, data acquired during rehabilitation are sequentially input as input parameters, and when a revelation is required, it is possible to output assistance predicted to be instructed. Action data that represents the action. That is, in the walking training apparatus 100 , it is possible to inform the training worker of the auxiliary action that is predicted to be revealed.

As can be seen from these examples, generally, the items and/or time ranges included in the second rehabilitation data described above differ depending on the learning model used in the learning unit 510b.

In addition, m output parameters (m is a positive integer) among the output parameters may be, for example, m setting values that exist for one of the setting parameters of the above (1). Similarly, one (1 is a positive integer) output parameter among the output parameters may be, for example, one detection timing or detection position that exists for one of the detection data in (2) above.

In these cases, the number of nodes in the output layer of the learned model increases. Therefore, a learned model can be constructed for each set parameter or detection data to be output, and a plurality of learned models can be constructed, such as a learned model for each help position to be output. Furthermore, by storing these learned models in the model storage unit 521 in advance, these learned models can be simultaneously used.

In addition, in the above example, it was demonstrated on the premise that the learning apparatus includes the level determination unit 510a, but the learning apparatus may not include the level determination unit 510a. In this case, the learning apparatus exemplified by the server 500 only needs to include an acquisition unit that acquires the determination result that determines the level indicating the evaluation of the training worker 901 among the determination results based on the first rehabilitation data . This acquisition unit can be constituted by, for example, the communication IF 514 and an acquisition control unit in the control unit 510 (for example, in the response processing unit 510 c ) that controls the communication IF 514 . The acquisition unit can be configured to acquire the determination result from a level determination unit provided in an external device such as a PC or the walking training device 100 . Alternatively, for example, a person may calculate the level based on the first rehabilitation data using a table calculation application software on a PC or the like. The acquisition unit in this case can be configured to input the calculated result (determination result) as input data.

Moreover, the case where the learning part 510b used the 2nd rehabilitation data corresponding to the training staff judged to be a predetermined level or more as teaching data to generate|occur|produce a learning completed model was demonstrated. This makes it possible to generate a learned model that takes into account the actions of the training assistants at or above a predetermined level.

On the other hand, as an alternative process, it is also possible to configure learning to be performed regardless of whether or not the level is equal to or higher than a predetermined level. For example, the learning unit 510b can generate a learning model by using, as teaching data, the plurality of levels tagged based on the determination results and the second rehabilitation data associated with the staff data corresponding to the plurality of levels. The processing of establishing the association here is equivalent to the preprocessing. The above-mentioned plurality of levels may be a plurality of levels as long as they are a part of all the levels to be determined, but may be all of the levels. By using such teaching data, it is possible to generate a learned model that takes the actions of the training staff into consideration for each level.

In other words, in the above-described substitution processing, first, a label of the level determined for the training worker is preliminarily added to each training worker (that is, for each worker data). Next, the learning unit 510b uses the second rehabilitation data (other than the staff data) and the staff data, that is, uses the second rehabilitation data including the staff data to associate with the labelled level and learns the second rehabilitation data. Action data included in health data.

For example, the higher the degree of excellence of the training staff, the higher the level is marked, and the higher the level of the label, the higher the weight in learning is associated with learning. If a more specific example is given, similar to an example in the case of using the second rehabilitation data of the training staff of a predetermined level or higher, the higher the level that can be determined, the more helpful the learning model will be given. It is realized by constructing a correct answer variable with a value such as (change of weight coefficient and threshold). However, in the above-mentioned alternative processing, the second rehabilitation data to be used is not limited to the data for training workers at a predetermined level or higher, as long as it is training work of a plurality of predetermined levels (preferably a plurality of consecutive levels). personnel data.

The learning unit 510b generates a learning model using the second rehabilitation data preprocessed based on the determination result as teaching data, as exemplified by the threshold processing based on the predetermined level and the alternative processing described above. In addition, the preprocessing here is not limited to the above-mentioned threshold value processing based on a predetermined level and the processing of associating with each level. For example, only the determination result may be associated with the second rehabilitation data. In any case, it is possible to generate a learning model that can suggest a preferred action to a training worker who assists the trainer when the trainer performs rehabilitation using the walking training device 100 .

(operation stage)

Next, the processing of the walking training apparatus 100 and the server 500 in the operation phase (inference phase) will be described. As described above, the walking training apparatus 100 can utilize the learned model by being configured to be able to access the learned model. In addition, the learned model can also be referred to as a learned module. In the operation stage, the walking training device 100 mainly cooperates with the server 500 connected to the network, that is, as a rehabilitation assistance system, to perform rehabilitation assistance processing.

In order to use the learned model as described above, the walking training device 100 can include the following output unit and notification unit. The output unit outputs the second rehabilitation data related to the rehabilitation performed by the trainer using the walking training apparatus 100 as input to the learned model, and can be exemplified by the input/output control unit 210c, the input/output unit 231, and the like. The above-mentioned notification unit notifies the training staff who assist the trainer in the rehabilitation of the action data output from the learned model, mainly by the notification control unit 210d, the display control unit 213, and the management monitor 139 (or the sound control unit). and speakers) etc.

On the other hand, on the server 500 side, the response processing unit 510c operates the learned model stored in the model storage unit 521 to perform response processing. In addition, the server 500 includes an input/output unit that inputs the second rehabilitation data output from the above-described output unit to the learned model and outputs the output from the learned model to the walking training device 100 . This input/output unit is exemplified by the communication IF 514 and the like.

Specifically, an example of rehabilitation assistance processing in the rehabilitation system including the server 500 will be described with reference to FIG. 6 . FIG. 6 is a flowchart for explaining an example of rehabilitation assistance processing in the server 500 .

First, the input/output control unit 210 c outputs the acquired data (second rehabilitation data) that may become an input parameter to the server 500 via the input/output unit 231 . The data acquired above may be data acquired at the start of rehabilitation, but may also be data acquired at various times during rehabilitation.

When the data is received via the communication IF 514 (in the case of YES in step S11 ), the response processing unit 510 c of the server 500 starts response processing. The response processing unit 510c analyzes the received data, divides it into a plurality of item data, and outputs the item data as each of the input parameters of the input layer in the learned model in the model storage unit 521 (step S12).

The response processing unit 510c operates the learned model to perform computation, and determines whether each output parameter from the output layer has an item of action data (an item indicating an auxiliary action) that needs to be revealed (notified) to the training staff. A determination is made (step S13). Here, each of the output parameters corresponds to each of the auxiliary actions to be notified. In addition, the determination of the output parameter can be performed by subjecting the value of the output parameter to threshold processing with a threshold value (or a common threshold value) prepared in advance for each. Of course, in the case of a model in which only two values of 0 and 1 exist for the value of the output parameter, it is only necessary to determine whether it is 0 or 1.

In the case of YES in step S13, the response processing unit 510c returns the information indicating the action data that needs to be notified (information indicating the item of auxiliary action) output from the learned model as an output parameter to the walking training device via the communication IF 514 100 side (step S14). The returned information can be an instruction to the walking training device 100 . When NO in step S13, the response processing unit 510c proceeds to step S15 described later without going through step S14.

In this way, in steps S13 and S14, the response processing unit 510c operates the learned model to execute an operation, and generates a command corresponding to the output parameter that is output as a value that needs to be revealed among the output parameters from the output layer. . On the other hand, the response processing unit 510c does not particularly process the parameters other than the above. That is, depending on the calculation result, the response processing unit 510c may not output a command at all, and this corresponds to a case where there is no need to inform (notify) the training staff. Here, the generation of the command can be performed, for example, by reading out the command corresponding to the output parameter from a pre-stored command group. In addition, the instruction may only express information indicating output parameters (eg, information indicating how many nodes are in the output layer). The response processing unit 510c transmits the generated command to the walking training device 100 side via the communication IF 514 .

After the process of step S14, the response processing unit 510c judges whether or not the reception of the second rehabilitation data is completed (step S15). Return to step S12.

In the walking training device 100, the input/output control unit 210c receives the command transmitted in step S14, and transfers it to the notification control unit 210d. The notification control unit 210d performs notification control corresponding to the command with respect to the display control unit 213, a sound control unit (not shown), or the like. The notification control unit 210d only needs to store notification controls corresponding to the command groups that may be transmitted from the server 500 side. The notification control unit 210d causes the display control unit 213 to output to the management monitor 139, for example, a display control signal for causing the management monitor 139 to display an image corresponding to the command. The notification control unit 210d causes the sound control unit to output to the speaker, for example, a sound control signal for causing the sound corresponding to the command to be output from the speaker. Among them, a part of the enlightenment, such as the enlightenment of the bare-handed assistance, may be the enlightenment based on the display of an image or a moving image explaining the method of the assistance.

Through such processing, the walking training apparatus 100 can use the acquired data as an input parameter, and output action data indicating an auxiliary action to be revealed (an auxiliary action performed by an excellent training staff) to inform the training staff. The auxiliary action. That is, in the walking training apparatus 100, the auxiliary action (setting, assistance, etc.) that should be performed next can be suggested based on such an enlightenment. In addition, since the learned model exists in the server 500 , it is possible to realize an operation using a common learned model for the plurality of walking training apparatuses 100 .

To give an example of use, for example, the walking training device 100 can be configured to input a data set composed of setting parameters set before the start of one rehabilitation into the learned model, and perform the exercise at the start of each rehabilitation as needed. Implications for setting parameters. For example, the walking training device 100 can be configured to use a data set composed of statistical values of data obtained during the above-mentioned predetermined period or rehabilitation during the above-mentioned predetermined period as an input, and may be configured to suggest set parameters and expected bare-hand assistance as needed .

The above has been described on the premise that output and notification are performed to training staff of all levels. This is because even good training staff have settings forgetting etc. to prevent this.

On the other hand, the walking training apparatus 100 can be configured to execute the notification-related processing only for the training staff 901 who cannot be said to be excellent, such as the need for notification. Specifically, first, the walking training apparatus 100 can be provided with a designation unit that designates, using a name, an ID, or the like, a training staff member 901 who assists the trainer during rehabilitation. This designation unit can be exemplified by, for example, a management monitor 139 including a touch sensor. Furthermore, the walking training device 100 can be configured to be accessible to a level storage unit that stores the level determined by the level determination unit 510a, in addition to the specifying unit. The horizontal storage unit may be, for example, a storage device within the overall control unit 210 or connected to the overall control unit 210 , or may be a storage device within the server 500 .

In addition, in the walking training apparatus 100, when the training worker 901 designated by the designation unit is not at a predetermined level or higher, the output unit outputs the second rehabilitation data, and the notification unit notifies it. That is, when the training staff 901 assisting the trainer 900 is a training staff of a predetermined level or higher, the walking training apparatus 100 in this example does not output the second rehabilitation data, and as a result, does not notify. Thereby, unnecessary notification is not given to the training staff who are supposed to not need notification.

Here, the description is made on the premise of threshold processing based on a predetermined level. However, it is not limited to this, and even in the case of the above-mentioned alternative processing, when the level of the training worker 901 specified by the specifying unit is the level used as the teaching data in the learned model, the walking training device 100 You only need to output and notify.

Next, with reference to FIG. 7 and FIG. 8, the example of notification to the training worker 901 in the above-mentioned walking training apparatus 100 is demonstrated. FIG. 8 is a diagram showing an example of an image presented to the training staff in the rehabilitation support process of FIG. 7 , and FIG. 9 is a diagram showing another example of such an image.

A GUI (Graphical User Interface) image 139a shown in FIG. 7 is an image in which a pop-up image 139b is superimposed on the image displayed on the management monitor 139 during rehabilitation. The pop-up image 139b is displayed when the walking training device 100 receives an instruction from the server 500 to perform an instruction to lower the walking speed by two levels. However, the image of the object on which the pop-up image 139b is superimposed is an image displayed at the time of enlightenment, and the content included in the image is arbitrary.

The GUI image 139c shown in FIG. 8 is an image in which a pop-up image 139d is superimposed on the image displayed on the management monitor 139 during rehabilitation. The pop-up image 139d is displayed when the walking training device 100 receives an instruction from the server 500 to raise the level of the swing assist by one level. However, the image of the object on which the pop-up image 139d is superimposed is an image displayed at the time of enlightenment, and the content included in the image is arbitrary.

(Effect)

As described above, in the learning apparatus according to the present embodiment, as the preparatory processing, the data related to the good training staff is classified based on the level distinction of the training staff, and the data related to the good training staff is used as the input to generate the learned model. The generated learned model can be a model that outputs good auxiliary actions (including changing the set value of the assist level, chatting, bare-handed assistance, etc.) as needed, and can also be a model that outputs good auxiliary actions when necessary. . Therefore, according to the present embodiment, it is possible to construct a learned model capable of outputting information indicating a good auxiliary action, that is, capable of revealing a preferable action to a training worker.

In addition, according to the walking training apparatus 100 according to the present embodiment, since the learned model generated in this way can be accessed, it is possible to suggest a preferred action to the training worker. Therefore, according to such a walking training device 100 , regardless of the degree of excellence of the training staff that may be generated based on years of experience, proficiency, ability, etc., it is possible to instruct the training staff in the same manner as in the case of assisting excellent training staff.

For example, when a forward propagation type neural network is used for the learning model, appropriate setting parameters and the like can be suggested as a response to the second rehabilitation data transmitted to the server 500 side before the rehabilitation starts. Similar to the case where the second rehabilitation data is periodically transmitted to the server 500 during rehabilitation, it is possible to receive enlightenment required at this time. For example, when a neural network having a recursive structure is used for the learning model, these implications can be predictively implemented in consideration of the previous second rehabilitation data. By making the statistics period, the number of storage steps, etc. appropriate for one data set, the timing of the display can also be made appropriate. In this way, in the walking training apparatus 100 according to the present embodiment, the change of the setting parameters, the execution of the chat, the execution of the bare-hand assistance, and the like can be informed at an appropriate timing.

(Supplement related to methods and procedures)

In the present embodiment, as can be seen from the above description, a learning method having the following acquisition steps and learning steps can also be provided. The obtaining step obtains output results of the output level, such as a judgment result of a level indicating a training worker's evaluation based on the first rehabilitation data. The learning step inputs second rehabilitation data including at least action data representing auxiliary actions performed by the training staff for the purpose of assisting the trainer, and generates a learning model that outputs action data for instructing the training staff's next action . In addition, the learning step generates a learning model using the second rehabilitation data preprocessed based on the output results such as the determination result as teaching data.

In the present embodiment, as can be seen from the above description, it is also possible to provide a rehabilitation assistance method in the walking training device 100 that can access the learned model obtained by the above-mentioned learning method, that is, the learned model (the walking training device 100 working method), which has an output step and a notification step as follows. In the output step, the walking training apparatus 100 outputs the second rehabilitation data related to the rehabilitation performed by the trainer using the walking training apparatus 100 as an input to the learned model. In the notification step, the walking training apparatus 100 notifies the training staff who assist the trainer during rehabilitation of the action data output from the learned model.

In the present embodiment, as can be seen from the above description, a program (learning program) for causing a computer to execute the above-mentioned acquisition step and learning step can also be provided. A learned model obtained by learning by a learning method, and a learned model obtained by learning by a learning program. In addition, in the present embodiment, as can be seen from the above description, a rehabilitation assistance program for causing the computer of the walking training apparatus 100 to be able to access the above-described learned model to execute the above-described output step and notification step can also be provided.

<Embodiment 2>

In the first embodiment, the server 500 is provided with the level determination unit 510a and the learning unit 510b, and the learned model is generated on the server 500. However, in this embodiment, the level output unit such as the level determination unit and the learning unit are provided in the The walking training device 100 side (for example, the overall control unit 210 ). The rehabilitation assistance system according to the present embodiment only needs to include the walking training device 100 . However, in this case, in order to increase the collection amount of rehabilitation data in the learning phase, it is preferable to configure such that rehabilitation data from other walking training apparatuses can be collected.

In addition, regarding the operation stage, an example in which the learned model is installed in the server 500 and the walking training device 100 transmits rehabilitation data to the server 500 and receives the action data is given, but the present invention is not limited to this. For example, the learned model can also be installed on the side of the walking training device 100 (eg, the storage unit in the overall control unit 210 ). Therefore, the walking training apparatus 100 can have a storage unit that stores the learned model. In addition, although not particularly described, in this embodiment, the various examples described in Embodiment 1 can be applied, and the same effects as those of Embodiment 1 can be obtained. As an example, in this embodiment, similarly to Embodiment 1, an acquisition unit may be provided instead of the level determination unit. That is, the walking training device 100 according to the present embodiment may include an acquisition unit instead of a level output unit such as a level determination unit.

<Embodiment 3>

Embodiment 3 will be described with reference to FIGS. 9 to 11 . 9 is a block diagram showing a configuration example of a server in the rehabilitation support system according to the third embodiment. The rehabilitation assistance system according to the present embodiment abbreviates its description, but can include rehabilitation assistance devices such as the walking training device 100 described in the first embodiment. In addition, although not particularly described, the various examples described in Embodiment 1 can also be applied to the present embodiment except for the following differences.

The learning apparatus according to the present embodiment is different from the learning apparatus according to Embodiment 1 in that it includes the following analysis unit instead of the level output unit such as the judgment unit exemplified by the level judgment unit 510a. The learning apparatus according to the present embodiment can be exemplified by the server 501, and the analysis unit can be the analysis unit 511a.

The server 501 shown in FIG. 9 can have a learning unit 511b and a response processing unit 511c corresponding to the learning unit 510b and the response processing unit 510c of the server 500 of FIG. 4 , respectively. The analysis unit 511a, the learning unit 511b, and the response processing unit 511c can be provided in the control unit 511 corresponding to the control unit 510 in FIG. 4 . The control unit 511 is basically provided with an analysis unit 511 a instead of the level determination unit 510 a in the control unit 510 . In particular, the response processing unit 511c can basically perform the same processing as the response processing unit 510c.

(Learning phase)

Next, processing in the learning phase of the control unit 511 of the server 501 will be described with reference to FIGS. 10 and 11 together. FIG. 10 is a schematic diagram showing an example of the result of the cluster analysis performed in the server, and FIG. 11 is a flowchart for explaining an example of the learning process in the server 501 .

The control unit 511 performs preprocessing on a part or all of the information included in the rehabilitation data, performs machine learning using the processed data, and constructs a learned model from an unlearned model. The analysis unit 511a executes preprocessing (preparatory processing), and the learning unit 511b executes machine learning. However, the control unit 511 can also be configured to execute preprocessing other than the processing in the analysis unit 511a at the same time.

First, the analysis unit 511a inputs the first rehabilitation data (step S21). The first rehabilitation data includes at least worker data representing the training worker 901 assisting the trainer 900 regarding the rehabilitation performed by the trainer 900 using the walking training apparatus 100 . In addition, the first rehabilitation data includes at least action data indicating the auxiliary action performed by the training staff 901 for the purpose of assisting the trainer 900 , and index data indicating the degree of recovery of the trainer 900 . In particular, since it is appropriate to judge whether or not a training worker is excellent, that is, whether it is an excellent training worker, the first rehabilitation data is appropriate based on the recovery index of the trainer, and therefore the index data is particularly important.

The analysis unit 511a performs cluster analysis on the above-described first rehabilitation data, and classifies the training staff (step S22). For the cluster analysis in the analysis unit 511a, for example, k-means can be used. Each cluster that is an analysis result is a result of classifying the trend of the first rehabilitation data, but is preferably adjusted to correspond to each data group classified according to the level of excellence of the training staff.

The cluster analysis in the analysis unit 511a can also use the extended k-means method and the X-means method (X-means) in which the designation of the number of clusters is performed automatically. In addition, various other techniques such as Gaussian Mixture Models (GMM: Gaussian Mixture Models) that can also obtain probability density distributions, and spectral clustering that focus on connectivity and perform clustering can be used for cluster analysis in the analysis unit 511a. In addition, in spectral clustering, data is first transformed into a graph, and in this transformation, an ε-nearest neighbor algorithm, a k-nearest neighbor algorithm (k-NN), a fully connected algorithm, and the like can be used.

In order to simplify the description, FIG. 10 shows an example of the result of cluster analysis for two parameters (two items) in the first rehabilitation data. In the example of FIG. 10 , the first rehabilitation data is classified into clusters C1 to C4 as a result of cluster analysis with the number of clusters (data groups) designated as four. In addition, since the number of parameters (the number of spatial axes) of the cluster analysis can usually be the number of items of the first rehabilitation data, it can be three or more in the case of the present embodiment.

The learning unit 511b inputs the second rehabilitation data including at least the action data, and generates a learned model that outputs the action data for indicating the next action of the training worker. In particular, the learning unit 511b selects the second rehabilitation data corresponding to the training staff included in one group (cluster) of the results classified by the analyzing unit 511a as the teaching data (step S23). Here, the learning unit 511b preferably uses the second rehabilitation data corresponding to the training staff included in only one group as the teaching data. Selection of teaching data will be described later.

Then, the learning unit 511b inputs the selected teaching data to the unlearned model to generate the learned model (step S24). In addition, the definition of each data in this embodiment, its preferred example, etc. are basically the same as those described in Embodiment 1, but the data selected as the teaching data may be different depending on the level determination unit 510a and the analysis unit 511a produce.

In addition, the learning unit 511b can use the second rehabilitation data corresponding to the training staff included in the group as teaching data for each of the plurality of groups in the result classified by the analysis unit 511a (classification result). That is, the learning unit 511b can be configured to generate a learned model using the second rehabilitation data of each of the plurality of groups as teaching data. Thereby, a plurality of learned models can be generated. In this case, the selection of the teaching data can be automatically performed by the learning unit 511b in a predetermined order or the like. In this case, the adjuster and the user of the learning model select and operate the learned model suitable for use. For the learned model, for example, a model that can be said to have a good correct answer rate from the viewpoints of the trainer's walking stability, FIM efficiency, walking speed, and physical ability can be selected as a model suitable for the specification.

In addition, the selection of the teaching data can be performed by an adjuster who adjusts the learning model. The adjuster can, for example, select a group that includes known good training staff. Therefore, the server 501 can be provided with a group designation unit that designates the above-mentioned group (cluster). However, the group designation unit may be configured to accept designation of clusters from an external terminal or the like. And the learning part 511b uses the 2nd rehabilitation data corresponding to the training worker included in the group designated by the group designation part as teaching data, and produces|generates a learned model. Thereby, the learned model of only the designated group can be generated.

In addition, in the above example, the description was made on the premise that the learning device includes the analysis unit 511a, but the learning device may not include the analysis unit 511a. In this case, the learning apparatus exemplified by the server 501 only needs to include an acquisition unit that acquires the classification results of the training workers classified into the first rehabilitation data by cluster analysis. This acquisition unit can be constituted by, for example, the communication IF 514 and an acquisition control unit in the control unit 511 (for example, in the response processing unit 511 c ) that controls the communication IF 514 . The acquisition unit can be configured to acquire the classification result from, for example, an analysis unit provided in an external device such as a PC or the walking training device 100 . Alternatively, for example, a person may use a cluster analysis application software on a PC or the like to execute a cluster analysis based on the first rehabilitation data. The acquisition unit in this case can be configured to input the result of its execution (the classification result, for example, the classified staff data) as input data.

In addition, the case where the learning unit 511b generates a learning model using the second rehabilitation data corresponding to the training staff included in one group of the classification results as teaching data has been described. This makes it possible to generate a learned model that takes into account the actions of the training staff members belonging to one group.

On the other hand, as an alternative process, the learning unit 511b may use, for example, a plurality of groups tagged based on the classification results and the second rehabilitation data associated with the staff data corresponding to each of the plurality of groups as the second rehabilitation data. Teaching data to generate learning models. The processing of establishing the association here is equivalent to the preprocessing. The plurality of groups described above may be a part of all the groups after classification, but may be all of the groups. By using such teaching data, it is possible to generate a learned model in which the actions of the training staff are considered in groups.

In other words, in the above-described replacement process, each classified group is first labeled. Next, the learning unit 511b uses the second rehabilitation data (excluding the staff data) and the staff data, that is, uses the second rehabilitation data including the staff data to tag the action data included in the second rehabilitation data with the tag. After the group establishes associations and learns. For example, the weights are labeled differently for each group, and the association of learning is established. For example, the labeling (labeling) can be performed in such a way that the weight is increased as the group including the particularly excellent training staff is increased, so that the staff data of any number of training staff with different degrees of excellence belong to which group the staff data belongs. The weights are different.

As described above, the learning unit 511b generates a learning model using the second rehabilitation data preprocessed based on the classification result as teaching data, as exemplified by the processing related to one group and the above-mentioned alternative processing. In addition, the preprocessing here is not limited to the above-mentioned processing related to one group and the processing of associating group by group, and for example, only the classification result may be associated with the second rehabilitation data. In any case, it is possible to generate a learning model that can suggest a preferred action to a training worker who assists the trainer when the trainer performs rehabilitation using the walking training device 100 .

(operation stage)

Next, processing in the operation phase in the walking training device 100 and the server 501 will be described. As described above, the walking training apparatus 100 is configured to be able to access the learned model, so that the learned model can be used. In the operation stage, the walking training device 100 mainly cooperates with the server 501 connected to the network, that is, as a rehabilitation assistance system, to perform rehabilitation assistance processing.

In order to use the learned model as described above, the walking training device 100 according to the present embodiment can include the output unit and the notification unit described in the first embodiment. Of course, the object to which the output unit in the present embodiment outputs the second rehabilitation data is the learned model generated in the present embodiment.

On the server 501 side, the response processing unit 511c operates the learned model stored in the model storage unit 521 to perform response processing. Furthermore, the server 501 has an input/output unit that inputs the second rehabilitation data output from the above-described output unit to the learned model and outputs the output from the learned model to the walking training device 100 . This input/output unit is exemplified by the communication IF 514 and the like. Such processing is basically the same as that described with reference to FIG. 6 , and the notification example thereof is also the same as that illustrated in FIGS. 7 and 8 .

Through such a process, the walking training device 100 can use the acquired data as an input parameter, output action data representing an auxiliary action to be revealed (an auxiliary action performed by an excellent training worker), and provide training information to the training. The staff revealed the auxiliary action. That is, in the walking training apparatus 100, the auxiliary action (setting, assistance, etc.) to be performed next can be suggested based on such an enlightenment.

In addition, the walking training apparatus 100 can include a designation unit that designates a training staff member who assists the trainer in the above-mentioned rehabilitation. This designation unit is the designation unit described in the first embodiment. In addition, the walking training device 100 can access the classification result storage unit that stores the analysis result (classification result) in the analysis unit 511a. The classification result storage unit may be, for example, a storage device within the overall control unit 210 or a storage device connected to the overall control unit 210 , but may also be a storage device within the server 501 .

Furthermore, in the walking training apparatus 100, when the training worker designated by the designating unit is a training worker who did not use the teaching data when generating the learned model, the output unit outputs the second rehabilitation data, and the notification unit outputs the second rehabilitation data. be notified. Therefore, for example, the analysis unit 511a can be configured to output, as a part of the analysis result, the name or ID of the training staff related to the first rehabilitation data serving as the teaching data. Thereby, unnecessary notification is not given to the training staff who are supposed to not need notification.

In addition, such output and notification are applicable also in the case of the above-mentioned alternative processing, and are not limited to the processing related to one group. That is, when the group to which the training worker 901 designated by the designation unit belongs is the group used as the teaching data in the learned model, the walking training device 100 only needs to output and notify.

(Effect)

Also in the present embodiment, as described above, the same effects as those in the first embodiment are exhibited. That is, in the walking training device 100 , it is possible to suggest to the training worker the auxiliary action (setting, assistance, etc.) that should be performed next.

(Supplement related to methods and procedures)

In the present embodiment, as can be seen from the above description, a learning method having the following acquisition steps and learning steps can also be provided. The acquiring step acquires a classification result obtained by classifying the training staff with respect to the first rehabilitation data by cluster analysis. The first rehabilitation data includes at least worker data on the rehabilitation performed by the trainer using the walking training apparatus 100 , action data indicating auxiliary actions performed by the training worker for the purpose of assisting the trainer, and data indicating the trainer's Indicator data for the degree of recovery. In the learning step, second rehabilitation data including at least action data is input, and a learning model that outputs action data for instructing the next action of the training worker is generated. In addition, the learning step generates a learning model using the second rehabilitation data preprocessed based on the classification result as teaching data.

In the present embodiment, as can be seen from the above description, it is also possible to provide a rehabilitation assistance method in the walking training device 100 that can access the learned model obtained by the above-mentioned learning method, that is, the learned model (the walking training device 100 work method). This method includes the output step and the notification step described in the first embodiment.

In the present embodiment, as can be seen from the above description, it is possible to provide a program (learning program) for causing a computer to execute the above-mentioned analysis step and learning step. Of course, it is also possible to provide a learned model obtained by learning by a learning device, A learned model obtained by learning by a learning method, and a learned model obtained by learning by a learning program. Further, in the present embodiment, as can be seen from the above description, it is also possible to provide a rehabilitation assistance program for causing the computer of the walking training apparatus 100 that has access to the above learned model to execute the above output step and notification step.

<Embodiment 4>

In the third embodiment, the server 501 includes the analysis unit 511a and the learning unit 511b, and the learned model is generated on the server 501, but in this embodiment, the analysis unit and the learning unit are provided on the walking training device 100 side ( For example, the overall control unit 210). The rehabilitation assistance system according to the present embodiment only needs to include the walking training device 100 . However, in this case, in order to increase the collection amount of rehabilitation data in the learning phase, it is preferable to configure such that rehabilitation data from other walking training apparatuses can be collected.

In addition, regarding the operation stage, an example in which the learned model is installed in the server 501 and the walking training device 100 transmits rehabilitation data to the server 501 and receives the action data is given, but the present invention is not limited to this. For example, the learned model can also be installed on the side of the walking training device 100 (eg, the storage unit in the overall control unit 210 ). Therefore, the walking training apparatus 100 can have a storage unit that stores the learned model. In addition, although not particularly described, the various examples described in Embodiments 1 and 3 can also be applied to this embodiment. As an example, in the present embodiment, similarly to the third embodiment, an acquisition unit may be provided instead of the analysis unit. That is, the walking training device 100 according to the present embodiment may include an acquisition unit instead of an analysis unit.

<Embodiment 5>

Embodiments 1 to 4 have been described on the premise that the notification is given to a person such as the training staff 901 , but it is also possible to notify a training assistant other than a human (a mechanical training assistant, that is, a human training assistant). Examples of the human training assistant include various human training assistants such as a humanoid robot, a voice assistant program, and a display assistant program. Taking an example in which the voice assistant program assists by voice, for example, "please lean your upper body further to the right", "please hold on to the handrail", "please reduce your walking speed", etc. can be spoken.

In the case where the training assistant is a program, it can be installed in the walking training apparatus 100 in an executable manner, but can also be installed in an executable manner in a mobile phone (also known as a smart phone) capable of communicating with the walking training apparatus 100 In the case of mobile phones), portable terminals such as mobile PCs, external servers, etc. In addition, the human training assistant can also have a program with artificial intelligence (AI program).

In addition, a plurality of human training assistants can be used in the walking training by the walking training apparatus 100, and can be managed separately from each other. That is, even when the training assistant is a human training assistant, the training assistant can be distinguished from other training assistants as in the case of the training staff.

When a human training assistant is used, the following data can be exemplified as the data (assistant data) related to the human training assistant corresponding to the data related to the training staff 901 in the above (4). For example, it is possible to cite the functions (audio assist function, video display assist function, etc.) possessed by the human training assistant (program), the name and version of the program, and the type that is continuously learned when the program is used. In the case of an AI program, a learning algorithm, degree of learning, learning time, number of times of learning, and the like can be cited.

In addition, when a plurality of training assistants (anything other than a human being) assist the rehabilitation at the same time, the rehabilitation data can include assistant data of a plurality of persons, as explained for the plurality of training staff. In addition, each assistant data can include information indicating whether it is a primary training assistant or an auxiliary training assistant. In addition to or instead of such information, each assistant data can include information indicating what kind of assistance is performed.

The notification in this embodiment will be described. For example, when it is necessary to notify a human training assistant other than the training staff 901, the notification control unit 210d only needs to notify the training assistant. The notification can be made directly through communication, but it can also be made using video and sound as in the case of a human, and a human training assistant can detect it. In addition, the human training assistant can change the settings of the walking training device 100 through communication, direct touch operation, or the like, so that even the human training assistant can execute the action suggested when the learned model is operated.

<Alternative example>

In each of the above-described embodiments, the example in which the trainer 900 represents a hemiplegic patient with a disease in one leg has been described, but the walking training apparatus 100 can also be applied to a patient with paralysis in both legs. In this case, the training is performed by wearing the walking assistance device 120 on both legs. In this case, the evaluation of abnormal walking can be performed for each diseased leg. The degree of recovery can be independently determined by evaluating the abnormal walking independently for each diseased leg.

In addition, although not shown in the figure, the walking training device 100 in FIG. 1 may not include the treadmill 131 , so that the trainer 900 can actually move in the space surrounded by the frame 130 . In this case, the frame 130 may be formed to be long in the traveling direction, and the protective belt stretch portion 112 , the front stretch portion 135 , and the rear stretch portion may be stretched along with the movement of the trainer 900 . 137 are respectively moved along the guide rails by a motor not shown. Since the trainer 900 actually moves relatively with respect to the floor surface, it is possible to further obtain a sense of achievement in the rehabilitation training. Of course, the walking training device is not limited to these configuration examples.

In addition, as described above, the rehabilitation assisting device according to each embodiment may be a device assisting other types of rehabilitation other than walking training or training other than rehabilitation. In this case, the learning device according to each embodiment can be a learning device that generates a learned model applied to the device, and can employ input parameters or output parameters corresponding to the type of rehabilitation or the type of training. Examples of training other than rehabilitation include exercise such as walking and running, training, and the like, and a training assistance device corresponding to the content of the training can be used. In addition, the index data in the case of training other than rehabilitation may be data indicating the degree of improvement of the physical function of the trainer instead of the degree of recovery of the trainer. As the degree of physical function improvement, improvement of muscle strength by exercise or the like, improvement of stamina, and the like can be included. In addition, even when the training is rehabilitation, the index data may be data indicating the degree of improvement of the physical function of the trainer, and in this case, the degree of improvement of the physical function can include the degree of recovery by rehabilitation or the like. In addition, in the case of training other than rehabilitation, the first rehabilitation data and the second rehabilitation data can be referred to as first training data and second training data, respectively, or simply referred to as first data and second data, respectively.

In addition, the rehabilitation assisting device described in each embodiment can also be constituted by a plurality of devices as a rehabilitation assisting system. Similarly, the walking training device can be configured by a plurality of devices as a walking training system, and the training assistance device can be configured by a plurality of devices as a training assistance system. In addition, the server (server device) described in each embodiment can include not only the learned model but not the learning device, but also all or part of the functions of the learning device, for example. In addition, the server device described in each embodiment can further include at least a part of the functions and parts described as the functions and parts of the rehabilitation assisting device. In addition, the above-mentioned rehabilitation assisting device or server device can be, for example, a hardware configuration including a processor, a memory, a communication interface, and the like. These means can be realized by a processor reading and executing a program stored in a memory.

Such a program, that is, the learning program and the learned model described in each embodiment will be described.

Such programs can be stored using various types of non-transitory computer readable media and supplied to a computer. The non-transitory computer-readable medium includes various types of tangible storage media. Examples of non-transitory computer-readable media include magnetic recording media (eg, floppy disks, magnetic tapes, hard drives), magneto-optical recording media (eg, magneto-optical disks). In addition, this example includes CD-ROM (Read Only Memory), CD-R, CD-R/W, and semiconductor memory. Examples of the semiconductor memory include mask ROM, PROM (Programmable ROM), EPROM (Erasable PROM), flash ROM, RAM (Random Access Memory), and the like. In addition, the program can be supplied to the computer through various types of transitory computer readable media. Examples of transitory computer-readable media include electrical signals, optical signals, and electromagnetic waves. The transitory computer-readable medium can supply the program to the computer through wired communication channels such as electric wires and optical fibers, or wireless communication channels.

From the above disclosure, it will be apparent that the embodiments of the present disclosure may be varied in various ways. These changes should not be considered as a departure from the spirit and scope of the present disclosure, and it is obvious to those skilled in the art that all such changes are intended to be included within the scope of the technical solution.

Claims (21)

1. A learning system is provided with:

an acquisition unit that acquires a classification result obtained by classifying a training assistant by cluster analysis with respect to 1 st rehabilitation data including at least assistant data indicating a training assistant that assists a trainer who has performed a rehabilitation exercise by a rehabilitation assistance system with respect to the trainer, action data indicating an assistance action performed by the training assistant for the purpose of assisting the trainer, and index data indicating a degree of recovery of the trainer; and

a learning unit that generates a learning model into which 2 nd rehabilitation data including at least the action data is input and outputs the action data for prompting a next action of the training assistant,

the learning unit generates the learning model by using the 2 nd rehabilitation data preprocessed based on the classification result as teaching data.

2. The learning system of claim 1 wherein,

the 2 nd rehabilitation data includes at least one of the index data and the assistant data.

3. The learning system according to claim 1 or 2, wherein,

the learning unit generates the learning model by using the 2 nd rehabilitation data corresponding to the training assistant included in 1 group of the classification results as teaching data.

4. The learning system according to claim 1 or 2, wherein,

the learning unit generates the learning model by using, as teaching data, a plurality of groups labeled based on the classification result and the 2 nd rehabilitation data associated with the assistant data corresponding to each of the plurality of groups.

5. The learning system according to any one of claims 1 to 4,

the learning system includes an analysis unit that performs the cluster analysis on the 1 st rehabilitation data to classify the training assistant,

the acquisition unit acquires a classification result obtained by classifying the training assistant from the analysis unit.

6. The learning system according to any one of claims 1 to 5,

the 1 st rehabilitation data and the 2 nd rehabilitation data include trainer data representing characteristics of the trainer.

7. The learning system of claim 6 wherein,

the trainer data includes symptom data indicating at least one of a disease and a symptom of the trainer.

8. The learning system according to any one of claims 1 to 7,

the action data includes at least one of data indicating an operation of changing a set value in the rehabilitation support system and data indicating a support action for the trainer.

9. The learning system of claim 8 wherein,

the data representing the operation includes data representing a proficiency of the operation.

10. The learning system according to any one of claims 1 to 3,

the learning unit generates the learning model by using, as teaching data, the 2 nd rehabilitation data corresponding to the training assistant included in the group for each of the plurality of groups in the classification result.

11. The learning system according to any one of claims 1 to 3,

the learning system is provided with a group specifying unit for specifying 1 group in the classification result,

the learning unit generates the learning model by using the 2 nd rehabilitation data corresponding to the training assistant included in the group specified by the group specifying unit as teaching data.

12. A learning system is provided with:

an acquisition unit that acquires a classification result obtained by classifying a training assistant by cluster analysis with respect to 1 st data including at least assistant data indicating a training assistant that assists a trainer who has performed training with a training assistance system with respect to the trainer, action data indicating an assistance action performed by the training assistant for the purpose of assisting the trainer, and index data indicating a physical function improvement level of the trainer; and

a learning unit that generates a learning model into which the 2 nd data including at least the action data is input and outputs the action data for prompting a next action of the training assistant,

the learning unit generates the learning model by using the 2 nd data preprocessed based on the classification result as teaching data.

13. A rehabilitation support system capable of accessing a learned model that is a learning model learned by the learning system according to any one of claims 1 to 11, the rehabilitation support system comprising:

an output unit that outputs the 2 nd rehabilitation data relating to the rehabilitation exercise performed by the trainer using the rehabilitation support system as an input to the learned model; and

a notification unit configured to notify the training assistant that assists the trainer in the rehabilitation exercise of the activity data output from the learned model.

14. A rehabilitation assistance system according to claim 13,

the rehabilitation support system is provided with a specification unit that specifies the training assistant that supports the trainer in the rehabilitation exercise,

the rehabilitation assisting system can access a classification result storage section that stores the classification result,

when the training assistant specified by the specifying unit is a training assistant that does not adopt the teaching data at the time of generation of the learned model, the output unit outputs the 2 nd rehabilitation data, and the notification unit notifies the user.

15. A computer-readable medium, wherein,

a learning model, i.e., a learned model obtained by learning with the learning system according to any one of claims 1 to 12 is recorded.

16. A learning method, wherein, having:

an acquisition step of acquiring a classification result obtained by classifying a training assistant by cluster analysis with respect to 1 st rehabilitation data including at least assistant data of the training assistant that assists a trainer who performs rehabilitation exercise with a rehabilitation assistance system, action data that represents an assistance action performed by the training assistant to assist the trainer, and index data that represents a degree of recovery of the trainer; and

a learning step of generating a learning model to which 2 nd rehabilitation data including at least the action data is input to output the action data for enlisting a next action of the training assistant,

the learning step generates the learning model using the 2 nd rehabilitation data preprocessed based on the classification result as teaching data.

17. A rehabilitation support method in a rehabilitation support system capable of accessing a learned model, which is a learning model obtained by learning by the learning method according to claim 16, the rehabilitation support method comprising:

an output step of outputting, by the rehabilitation support system, the 2 nd rehabilitation data relating to a rehabilitation exercise performed by a trainer using the rehabilitation support system as an input to the learned model; and

a notifying step of notifying, by the rehabilitation assistance system, the exercise assistant that assists the trainer in the rehabilitation exercise of the action data output from the learned model.

18. A computer-readable medium, wherein,

a learning model, i.e., a learned model obtained by learning by the learning method according to claim 16 is recorded.

19. A computer-readable medium in which a program for causing a computer to execute:

an acquisition step of acquiring a classification result obtained by classifying a training assistant by cluster analysis with respect to 1 st rehabilitation data including at least assistant data of the training assistant that assists a trainer who performs rehabilitation exercise with a rehabilitation assistance system, action data that represents an assistance action performed by the training assistant to assist the trainer, and index data that represents a degree of recovery of the trainer; and

a learning step of generating a learning model to which 2 nd rehabilitation data including at least the action data is input to output the action data for enlisting a next action of the training assistant,

the learning step generates the learning model using the 2 nd rehabilitation data preprocessed based on the classification result as teaching data.

20. A computer-readable medium, wherein,

a rehabilitation support program for causing a computer having access to a rehabilitation support system of a learned model, which is a learning model learned by the program recorded in the computer-readable medium according to claim 19, to execute:

an output step of outputting the 2 nd rehabilitation data relating to the rehabilitation exercise performed by the trainer using the rehabilitation support system as an input to the learned model; and

a notifying step of notifying the training assistant that assists the trainer in the rehabilitation exercise of the action data output from the learned model.

21. A computer-readable medium, wherein,

a learning model, i.e., a learned model learned by the program recorded on the computer-readable medium of claim 19 is recorded.

Images